A Learning Technique Inspired by Computer Hackers

Malicious hackers commonly use a technique called the 'man-in-the-middle attack'. In that attack, the bad guy interposes himself between two people who are trying to communicate, eavesdrops on their communications, and potentially alters their messages. For example, let's say Adam is hanging out at a coffee shop when he remembers that he needs to email detailed wire transfer instructions to his new investor, Barbara. If Chuck hacks into the wireless router, he can intercept Adam's email, insert his own bank account details, and forward the altered email on to Barbara. Since Adam and Barbara don't know there's someone monitoring and modifying their communications, they won't realize they're being duped until Barbara's money is in Chuck's bank account.

Hackers use the man-in-the-middle attack for evil, but you can use it for good whenever you introduce two people to each other. The type of introductions I'm referring to are where one person with a specific question is introduced to another person who might be able to answer that question. For example: 

  • "Oh, you're interested in financial advice? You should meet my friend Roger."
  • "My friend Tammy is the best growth hacker I know. She can help you figure out how to make your mobile app more viral."
  • "If you want recommendations for things in to do in Spain, talk to Amy and Luke -- they spent 2 months there last summer."

These introductions are useful for the people who need help, but they also present a golden opportunity to learn something interesting. Here are two man-in-the-middle approaches that I've personally used to learn about a variety of topics:

Listen and Repeat

Sometimes you can circumvent an introduction by forwarding a question from person 1 to person 2, then passing person 2's answer back to person 1. This works well for questions that don't require a lot of additional context.

Example

Before:

"John, meet Jane. Jane can help you with setting up an accounting system for your small business."

After:

Step 1: Email John -- "Hey John, what kind of questions do you have about setting up an accounting system?"

Step 2: Email Jane -- "Hey Jane, my friend John is setting up a small business and has the following questions about accounting. Do you have any advice that I can pass along to him?"

Step 3: Email John -- "Hey John, one of my friends is an accountant and I sent her the questions that you had. She said..."

A Fly on the Wall

Introduce two people and ask if you can sit in on their discussion. This works well for open-ended topics where you won't be able to provide enough context to use the "Listen and Repeat" approach.

Example

Before:

"John, meet Jane. Jane can help you with setting up an accounting system for your small business."

After:

"John, meet Jane. Jane can help you with setting up an accounting system for your small business. Once you two decide on a time to talk, I'd love to join you if you don't mind. I, too, am interested in learning about accounting systems."


The next time that you're about to introduce two people, think about whether the topic they'll be discussing is personally interesting to you. If it is, turn the introduction into a learning opportunity.

Extracting More Value from Investors

Investors can be a great resource for founders -- and not just because of the capital they provide. Many investors have great operating experience and vast networks, or they've at least observed many startups from the sidelines and can offer good advice based on the patterns they've seen. Unfortunately, a lot of founders underutilize their investor network.

Here are the three most common mistakes that I see:

  • Never asking for help and declining offers of help. I'm not sure if this happens because founders feel guilty about asking for help, or if they're afraid of being judged negatively if they aren't handling everything by themselves, or if they think investors won't be of much help. The reality is that most investors like being helpful, and they have a lot to offer. They invested in your company because they'd like to work with you and because they hope to get a good ROI, and being helpful addresses both desires. Furthermore, any advice you get is not binding, so if you don't like it, you don't have to use it. It's still your company to run.
  • Asking for the same things as every other company. Every month, I see 10+ investor update emails that include some variant of "We need help finding awesome engineers." I know. I really want to help -- and sometimes I can -- but it's hard when every single company is also looking for awesome engineers. I might meet 1-2 strong candidates each month, but if there are 30 companies vying for those candidates, the math just doesn't work out. A more reliable way to get help is to ask for things that are not zero-sum. For example: the opportunity to talk with a cryptography expert for an hour, an introduction to someone on Google's Chrome team, feedback on your website redesign, and so on.
  • Only asking for help in one or two areas. It's actually perfectly fine to ask for help with hiring if that's not the only thing you ask for. The problem is that many founders ask for that and nothing else. By offering more options, you maximize the chances of each investor being able to help you somehow.

Fundamentally, most investors can be viewed as 1-2 hours of free labor every month. There are exceptions, like board members/major investors who'll be willing to put in a lot more time, or very casual investors who might not want to invest any time at all, but most investors are good for at least some (free) work every month.

So what can you ask investors for? Almost anything! Some ideas:

Introductions to...

  • Potential hires.
  • Specific customers or partners. ("Looking for an intro to a Director of Eng or higher at Foursquare.")
  • Customers in a specific vertical. ("Looking for intros to mobile gaming companies with at least $2m in annual revenues.")
  • Experts that could help your company. ("Looking for info on handling security questions from Fortune 500 companies.")

Comps for...

  • Leasing office space. ("What is a typical price per employee for SoMa/downtown Palo Alto/etc?")
  • Equity and salary grants. ("What is a typical comp package for a VP of Sales who is employee #8?")
  • Commission structures and goals for salespeople. ("What's a good monthly sales quota for a product with our target customers and price point?")
  • Company metrics. ("What is a good goal for my churn rate?" or "What is a typical CTR for marketing emails?")

Feedback on..

  • Website UI, UX, copy, etc.
  • Mobile app UI/UX.
  • Sales decks, pitch decks, marketing materials, etc.
  • Resumes of people you're thinking of hiring.

Advice for...

  • How to make the most of various growth channels. ("What are some best practices for using Facebook Ads?")
  • How to approach your next round of financing. 
  • Which vendors to use for PR, SEO, Health/Dental benefits, etc.
  • General company strategy: product roadmaps, expansion plans, customer segments to target, etc.

So next time you write an investor update email (and you should be writing those monthly!), list out 5-10 things that you'd like help with. It takes very little effort, and you might be surprised by the results.

Analyzing AngelList Job Postings, Part 2: Salary and Equity Benchmarks

A few weeks ago, I did a basic analysis of AngelList job postings. That analysis looked at attributes like job locations, vesting schedules, and commonly requested skills.

In this post, I'll look at salary and equity numbers based on startup size. In my experience, founders frequently give out equity grants that are too generous or too stingy. The dangers of stinginess are that candidates will choose to work somewhere else and that you will waste a lot of time interviewing people who definitely won't take your offer. The danger of generosity is that, especially for the first few hires, you are giving away much more equity than you need to. That equity could be used to give stronger offers to multiple candidates later on, to raise more money from investors, or to retain more decision-making power for founders.

Background

Every job posting on AngelList has a salary range and an equity range. For example, at the time of this writing, Casetext is hiring a full-stack software engineer and offering $90k - $125k in salary and 0.3% - 1.5% in equity.

Two notes on these ranges: first, the more senior and well-qualified you are, the closer your offer will be to the top of the company's salary/equity ranges. Second, a lot of companies will let you trade equity for salary. That is, if a company is offering 1% - 2% equity and $100k - $130k salary, a specific candidate might get to choose between $100k and 1.75% or $125k and 1.2%.

Methodology

I used AngelList's API to find all full-time jobs in Silicon Valley that were posted in the last 2 months and offered salaries of at least $20k. I then checked LinkedIn and manually counted the number of employees at each company. (This was exactly as fun as it sounds.) To keep the manual work manageable, I limited the companies I was looking at to those with an AngelList Signal score of 6 or more, which cut the number of companies in half. All of these filters resulted in a list of jobs at over 300 startups. Finally, I grouped startups by the number of employees they had, then graphed the equity and salary offers for different company sizes.

Caveats

My methodology is far from perfect, and there are lots of areas where fuzziness was introduced. For example, not everyone is on LinkedIn, so employee counts based on LinkedIn data won't be perfectly accurate. The following data is meant to provide ballpark ranges for salary and equity offers. Please don't treat it as gospel.

Summary of Results

Note: These are benchmarks based on a medium-sized sample; they're not hard-and-fast rules. The benchmarks are for engineering jobs in Silicon Valley. (Non-engineering jobs are discussed at the end of this post.)

Assuming that a startup has two founders, here are some ballpark numbers for engineering job offers:

Salary:

  • For employee #1:
    • 20th percentile salary range is $70k - $100k
    • 50th percentile salary range is $80k - $120k
    • 80th percentile salary range is $82k - $135k
  • For employees #2 through #13, salaries rise for higher paying jobs:
    • 20th percentile salary range is $75k - $100k
    • 50th percentile salary range is $85k - $125k
    • 80th percentile salary range is $100k - $150k
  • For employees #14 through #35, salaries rise for lower paying jobs :
    • 20th percentile salary range is $78k - $120k
    • 50th percentile salary range is $90k - $134k
    • 80th percentile salary range is $102k - $150k
Lower salary ranges tend to be for more junior or more specialized roles (e.g. front-end engineer). Higher salary ranges tend to be for more senior roles, or for full-stack engineers. That is, a Junior Front-End Engineer is more likely to get an offer in the 20th percentile, while a Senior Full-Stack Engineer is more likely to get an offer in the 80th percentile.

Equity:

  • Hire #1: 2% - 3% of equity
  • Hires #2 through #5: 1% - 2%
  • Hires #6 and #7: 0.5% - 1%
  • Hires #8 through #14: 0.4% - 0.8%
  • Hires #15 through #19: 0.3% - 0.7%
  • Hires #21 through #27: 0.25% - 0.6%
  • Hires #28 through #34: 0.25% - 0.5%

These ranges indicate the maximum equity amounts offered by companies. These amounts are typically reserved for ideal candidates, or candidates who are very senior. If you're just out of college or not a perfect fit for a company, you will probably get a little (or a lot) less. A good way to frame these numbers is to add "up to" before each range. For example, a typical 6th hire will get up to 0.5%-1%.


Given the recent attention to burn rates, one interesting observation here is that the first ten employees of a company will own about 12% of the company, and will cost, on average, about $100k/month for salary alone. Companies that are able to raise $1.5m+ seed rounds are fortunate because they can usually get about 18 months of runway as their team grows from a few founders to 10-15 people.

Salaries for Engineers

After employee #1, salaries for employees were relatively close together, with one noticeable jump occurring around the 13th or 14th hire. I suspect this is around the size of a company raising a Series A, and as soon as an A round is closed, the company increases salaries a little bit.

Here are graphs of min/max engineering salaries:

The way to interpret these graphs is that the horizontal axis is the salary's percentile among all other salaries. For example, if a company is offering a salary range of $90k - $140k for an engineer who will be the 8th employee, then we can see that $90k and $140k are each at the 60th-70th percentile on their respective graphs, so the salary offer is pretty good (better than 2/3 of startups).

    Equity for Engineers

    The following graphs are histograms for max equity stakes of different job offers. Each bar represents a single startup's offer. For example, the first graph shows that for hire #1, 6 startups offered up to 2% equity, 4 startups offered up to 3%, and 3 startups offered up to 5%.


    I expected to see a consistent drop for each additional employee, but was surprised to find that offers remained consistently high for many employees at a time. For example, the equity for Hire #5 is often similar to the equity for Hire #2. This feels like a market inefficiency, as Hire #5 is taking on much less risk than Hire #2.

    What about latecomer VPs, Directors, etc?

    The dataset only had 7 data points for VPs and Directors of Engineering who came in when a company already had 10+ people. In those cases, their equity stakes were at the top of the ranges for their employee numbers (e.g. 1.5% for a VP who was employee #13), and their salary ranges were consistently around $120k - $180k.

    What about non-Engineers?

    The sample size of non-engineering jobs at early stage companies is too small to analyze rigorously. That said, here are some high-level observations:

    Sales jobs

    VPs tend to get 1%-2%. Directors tend to get 0.5% - 1.0%. Salary ranges for directors are often $80k - $140k.

    Non-director role salaries were very inconsistent, ranging from $40k - 60k to $80k - $120k (for the same job titles). Presumably there's also a commission structure, but that's not captured by AngelList.

    Equity stakes for non-director roles were generally 1% or less. For the first 10 hires, equity was typically 0.3% - 0.5%, then dropped off to 0.1% - 0.2% for subsequent hires.

    Marketing jobs

    VPs: not enough data

    Directors at < 15-person companies: $80k-$120k, 0.5% - 1.0%

    Directors at >= 15 person companies: $120k-$180k, 0.25% - 0.5%

    UI/UX/Designer jobs

    Median salary range was around $80k-$130k.

    Designers among the first four hires get up to 1-2% equity, occasionally only 0.5%. Designers among next 5 hires get up to 0.5% - 1.0%. 0.2% - 0.5% for employees #10-30.

    I'd like to reiterate that there wasn't enough data to deeply analyze non-engineering jobs. There were sales reps who joined early and only got 0.05% stakes, as well as marketers who joined late and got 1% stakes. 

    What about outside of Silicon Valley?

    I chose to focus on Silicon Valley because the dataset was more manageable. For other locales like NYC or Houston, a hack that might work is to look at a dozen random, high-quality companies and see how their offers compare to Silicon Valley's. Seeing where those offers fall can help calibrate expectations.

    That's a lot of data. Now what?

    There are four common problems with startup job offers:

    1. The founder's offer is too generous. In this case, the employee will be happy in the short term, but the company suffers in the long term. If the founder gives away 10% more than they should to the first 10 employees, then that's 10% less that's available for fundraising and future equity grants -- and that might break the company.
    2. The founder's offer is too stingy. In this case, the entire interview process is usually a waste of time for both parties. Furthermore, even if an employee takes the offer, it's harder to retain them because they'll either resent the founder once they realize they're being paid below market, or they'll just move to a company that pays better.
    3. The employee's expectations are too high. In this case, the interview process is likely a waste of time, and the employee is likely to pass on multiple good job opportunities before realizing they have unrealistic expectations.
    4. The employee's expectations are too low. If a founder exploits this, it hurts long-term retention and builds long-term resentment among employees.

    The goal of this analysis is to increase transparency, which can help mitigate these four issues. The benchmarks are just rough guidelines, but they can be helpful if the employee expects to receive 2% while the founder thinks 0.2% is fair, or vice versa.

    As a parting thought, I'd like to reiterate that these are not rules, they are starting points. There are plenty of reasons to have higher equity grants (e.g. a specific employee can change the entire direction of a company) and plenty of reasons for lower grants (e.g. there are significant non-financial perks, like working with an amazing founder or working on a particularly meaningful project).


    In the next and final post in this series, I'm going to look at which aspects of a company and job correspond to getting more job applicants via AngelList. If you'd like to be notified when that post is published, please subscribe to this blog or check back in a week or two.

    Thanks to Mike Greenfield and Cheng-Tao Chu for feedback on earlier drafts of this post.

    Extensive Notes from SalesConf

    As an engineer with no direct sales experience, I've always wanted to learn more about how to turn code into cash, but wasn't sure where to start. A few months ago, I got an invite to SalesConf, an affordable sales conference with a great speaker lineup. I bit the bullet and signed up, and I'm happy to say that I learned a ton. While the conference was targeted at SaaS startups, many of the lessons are broadly applicable.

    This post is a collection of my notes from the conference, which was held last Friday.

    TL;DR

    There are a lot of notes here -- about 6,000 words in all -- and not everyone has the time to read that much. While I think there's a lot of value to reading the entire post, here is a summary of high-level points that came up in multiple talks:

    • You need a sales process -- don't just wing it, or expect your initial sales employees to just figure things out on their own.
    • Your pitch should be simple and clear, and tailored to the prospective customer you're talking to.
    • Know what you're looking for in an ideal customer and an ideal salesperson. 
    • Measure everything. Analyze the data for your best customers and look for patterns. Go out and find more customers who fit those patterns.
    • Specialize and do 1-2 things very well instead of doing a lot of things poorly. This advice applies both to team efforts and to individual efforts.
    • There are many great software tools for salespeople and marketers (links throughout the notes). Use them.
    • When your company is small, you don't have the bandwidth to handle a lot of potential customers, so identify your ideal customer profile and focus on companies that fit that profile.
    • Ask customers to pay for up-front development (for new products) and custom development (for enterprise PoCs).
    • Your first sales hire needs to be a great salesperson and to be able to build out a team. Don't settle for just one of those two skills.
    • Understand the milestones that lead to sales and customer success, then work hard to lead customers to those milestones.

    Ok, on to the notes! If something doesn't make sense, it's my fault and not the speakers'.

    Table of Contents

    1. How to Crush Your Sales Goals by Aaron Ross
    2. Uncovering a Treasure Trove of Sales Opportunities With Customer Data by Lincoln Murphy
    3. How To Fast Track Sales Reps To Peak Performance by Bridget Gleason
    4. Step-By-Step: $290K MRR In 14 Months by Tim Sae Koo
    5. Build A Massive Personal Brand-Powered Referral Network by Carolyn Betts
    6. Build a Hyper Targeted Lead List in 12 Minutes by Ilya Lichtenstein
    7. How to Sell a Product Before You Create It by Josh Isaak
    8. Growing B2B and SaaS Sales Teams by Armando Mann
    9. How to Write a 3m Page View Article by Nick O'Neill
    10. How Mattermark Sold Their First Million by Danielle Morrill
    11. Sales Hot Seats with Hiten Shah, Lincoln Murphy, and Steli Efti

    How to Crush Your Sales Goals by Aaron Ross

    Speaker background: Aaron spent 2002-2006 at Salesforce, where he built out the outbound prospecting team. He is also the author of Predictable Revenue. [affiliate link]

    Aaron talked about common sales mistakes and how to fix them. 

    Fatal mistake #1: not having a sales system

    Indicators of this mistake include high (>10% per year) churn on the sales team and/or missing your goals. You need to have a process in place, be able to describe your ideal customer, etc.

    Corollary: at large companies, missing quotas is often blamed on salespeople, who are then fired and replaced. But if recruiting, training, and other processes remain the same and quota continue to be missed, the problem is with the processes, not the salespeople.

    Fatal mistake #2: confusing your prospects

    When people are confused by your pitch, they default to saying "no." Don't pitch your product as a kitchen sink or ask your prospective customers what features they want. Instead, try to understand each customer's problems, then give a very focused, targeted pitch which makes it easy for them to make a yes/no decision. Confusion always defaults to a 'no' reply.

    Make the pitch about what customers want. They don't care about what you do, they care about what you can do for them. For example, don't talk about your "scalable platform," because a prospect won't care; they only care that you can solve their specific problems.

    Tips for improving messaging:

    • Instead of telling someone what you do from your point of view, pretend they asked you: "How do you help customers?"
    • Review your own pitch. For each bullet point, ask yourself, "so what?" or "what's so great about that?" Your answers are what you should be pitching to prospects.
    • Selling ideas is better than selling benefits, which is much better than selling features. Aaron used a great analogy:
      • Selling a feature: "This is a drill."
      • Selling a benefit: "You need to put a hole in the wall, and this drill makes that possible."
      • Selling an idea: "You'll be happier if you can cover your walls with family photos, and this drill makes that possible."

    Fatal mistake #3: driving growth by growing the sales team

    It's less common today, but the way people used to think about growing sales was "I have 10 salespeople, and I need to double sales, so I'll double the number of salespeople." The real lever is lead gen. If you have great sales but terrible lead gen, you'll struggle; if you have great lead gen, you can mess up the sales process and still succeed.

    3 types of leads: seeds (word of mouth), nets (marketing), and spears (outbound sales).

    Tips for word of mouth:

    • Improve referral rates by hiring customer success reps. Happy customers will recommend you to others.
    • You should hire a customer success manager (CSM) by the time you have 5-15 people, and should have one CSM for every $1m-$2m in annual revenue. 
    • Two baseline metrics: customer churn should be <15%, revenue churn should not be negative (if you lose customers, you win more revenue from upselling to existing customers and growing your customer base)
    • Triggers for having a CSM engage a customer: support/help desk interactions, billing/payment history (i.e. someone is paying on time), survey feedback, engagement with marketing materials, levels of usage of the product and specific features.
    • GainSight offers great dashboards for customer success management.

    Tips for marketing:

    • Just like the VP of Sales has a sales quota, the VP of Marketing should have a lead quota (with some balance of quantity vs. quality).
    • Most important metric to track: lead velocity rate. This is a measure of how much qualified pipeline is being created each month and helps you estimate growth/company health. Ideally, the velocity is high and growing every month.
    • Make your marketing emails personal. Messaging that sounds more human gets much better results. For example, use text emails instead of HTML with a lot images, and include a one-sentence story about yourself (even if it's something trivial like "I just finished my coffee and wanted to send this to you.") This advice is probably simplest thing you can do to make your marketing better.

    Tips of outbound prospecting (Aaron's specialty):

    • Your funnel should be to identify target customers, prospect with cold emails/calls/appointments, then start the sales cycle on qualified leads. For example, 1000 emails => 100 calls => 20 appointments => 15 sales qualified leads, each worth $50k/year.
    • Common sales email fails: too long, confusing/lots of jargon, filled with lies, boring, vague calls to action. Aaron's mantra: "make it simple to understand and easy to answer."
    • If you're vague, like "let me know the best way to reach you," it's easy for a recipient to not respond. Simple, specific questions are much more likely to get answers. For example, "Are you free Thursday at 3pm for a phone call?" Even if the reply is 'no', at least you've gotten the prospect talking.
    • Keep it short. Relevant HubSpot study result: the highest response rates were for emails that are 300-500 characters (1-4 sentences). Keep emails personal, but take out filler material. Don't beat around the bush.

    Fatal mistake #4: not specializing and trying to do too much

    Salespeople often avoid prospecting because they're either not good at it or don't like it.

    Over time, you want inbound and outbound salespeople who qualify leads, account execs who close sales, and customer success managers who increase the value of existing accounts.

    If salespeople own the entire process, customers get horrible service because salespeople focus on deals they're trying to close that month, and prospecting and customer success suffer.

    Even if you're at a small startup and you own the entire sales process, at least separate functions by day of the week so that you're not neglecting anything (e.g. Tuesdays are for prospecting; Thursdays are for customer success.)

    Most marketers try to do too much (webinars, events, white papers, etc). It's better to do a few things very well instead of a lot of things poorly.

    Specialization leads to predictability: you will get insights on what works and what doesn't, better scalability for each layer, and a talent/farm team system where you can hire more junior people and help them grow into more impactful roles. You will struggle without specialization.

    For some parts of the process, outsourcing can be effective if done right. See: LeadGenius and Carburetor (Aaron's company).

    Salesperson churn is usually due to unrealistic quotas, poor management, or lack of coaching. It's rarely about compensation.

    Online resources that Aaron recommended during the talk:

    Uncovering a Treasure Trove of Sales Opportunities With Customer Data by Lincoln Murphy

    Speaker background: Lincoln is a customer success evangelist at Gainsight, which offers products for customer success management. He also has a great blog called SaaS Growth Strategies.

    Tip #1: Even if you don't have enough customers for meaningful data right now, you need to start collecting now.

    Tip #2: Focus on your ideal customer. Founders are often afraid to do this because they have a fear of missing out on other customers, but in the early days, you don't have the resources to handle everyone, so you may as well be choosy.

    Ideal customers have 7 key attributes: ready to buy, willing to buy, able to buy, will get value/success out of using your product, profitable (you can make money off of selling to them), have expansion/upsell potential, and have advocacy potential. Lincoln focused on the last 4 attributes during his talk.

    General methodology for finding ideal customers:

    1. Determine what situation you're solving for (e.g. potential customers who will refer others or sign a contract quickly or pay a lot.)
    2. Get access to your customer data.
    3. Find existing customers who match your situational criteria.
    4. Look at all of the data you have for those customers: which product features do they use? What industries are they in? How many employees do they have?
    5. Look for patterns among customers who match the profile you're looking for.
    6. Use those patterns to find new customers who match your profile.
    7. Profit!

    Basically, find your most successful customers, then try to find more customers who are just like them.

    Different types of customers, and the patterns that reveal them:

    • Customers who are most likely to find success with your product. Data to look at: average Net Promoter scores over time, short sales cycle length, quick time to getting value out of your product, customer support data (tickets that are resolved quickly and thoroughly). You can also poll your account managers for their gut "health" scores for each customer. That may not sound scientific, but it's basically wisdom of crowds and it works. (Gainsight, Lincoln's company, makes tracking all of this stuff easy.)
    • Customers who are the most profitable. Data to look at: on the cost side, look at fully-loaded customer acquisition cost (ads, sales time, trials and POCs, cost of converting from free to paid). Include post-sale costs and ongoing support costs. On the revenue side, look for high lifetime value, long term contracts, increasing usage/purchasing of add-ons over time, and low support costs. Also look for short sales cycles.
    • Customers whose accounts grow over time. Data to look at: intra-company virality (land and expand), roles/departments of initial purchasers within the company.
    • Customers who are your biggest advocates. Look for: companies that speak at your events, participate in your case studies, refer customers directly, provide testimonials, etc.

    In general, customer success is a prerequisite for each of these customer types. That is, if customers are not getting value out of your product, they are not going to increase their spending, advocate for you, or be profitable for very long.

    How To Fast Track Sales Reps To Peak Performance by Bridget Gleason

    Speaker background: Bridget is VP of Sales at Yesware, which offers email-based tools that make salespeople more efficient and more effective. 

    Everyone wants to be a peak performer, but how do you increase each person's probability of success?

    • Culture always precedes great results. If people don't hit numbers and the rest of the company is okay with that, you get a culture of mediocrity. (Personal note: the reminds me of the Broken Window Theory in software development.)
    • Need to hire A players. If the team isn't exceptional, then other exceptional people won't want to join.
    • Need a great system. "A great assembly line comes before a high quality car."

    5 things to think about as you create your system and assemble your team: 

    Ideal sales reps

    Just like you expect salespeople to know ideal customer profiles, you should know the ideal sales rep profile (which might change over time). For your first hire, the ideal rep is probably someone who is both a great individual contributor and a great team builder -- not just one of the two.

    Strong onboarding program

    You can't just hire people and then expect them to hit the ground running. You should have a training plan for making reps more successful. This plan can include content you create, or a collection of public content like blog posts and videos. Regardless, you shouldn't expect people to come in and just 'figure it out'. You don't have to teach people everything at the beginning, but you need to provide enough for them to get started.

    Clear expectations

    New reps should know what to expect in terms of what they'll be working on, their works hours, quotas and goals, attitudes, etc.

    Team involvement

    Involve the rest of your team in onboarding (via training, mentorship, social events, etc). The more the whole team is involved, the more everyone feels vested in the entire team's success.

    Commitment to the system

    Onboarding is not a one time event, it's an ongoing commitment.

    Q&A

    Q: Thoughts on commission-only vs. commission + salary?

    A: Not a fan of commission-only because people aren't as committed. In the early days, when there's less predictability in the sales process, might need to pay higher base and less commission. Over time, shift to lower base and higher commission. Bridget favors people who prefer more commission because they are betting on themselves. At her previous company, Engine Yard, new hires were given a choice of taking more salary or more commission. (Personal note: this is like engineers often having a choice between more salary or more equity.)

    Step-By-Step: $290K MRR In 14 Months by Tim Sae Koo

    Speaker background: Tim is the CEO of Tint, which lets you display social feeds anywhere.

    Path to $290k MRR (and lessons learned along the way):

    • First 10 customers: LinkedIn direct targeting. Used Rapportive hack to guess emails.
    • Change your LinkedIn settings to let people see when you've viewed them. Occasionally, people you view might contact you.
    • Charging early makes you seem valuable from the get-go. Release a beta version of your product and your business model.
    • Charge more. People are afraid of charging because they want growth and usage. However, charging more lets you quickly figure out who your real customers are, the ideal customer profile, etc.
    • Charge based on value, not cost. Ask customers, "if you built this yourself, how much would it cost?" Charge based on their answers. Tim used Qualaroo on the pricing page to learn about what customers wanted.
    • You need a distribution strategy. Ask: where do your customers look for tools? In this case, Tint integrated with app stores, built WordPress plugins, etc.
    • Work on SEO for terms your customers search for. Add backlinks to landing pages within your product -- especially for non-premium customers. Those customers may not pay, but at least they will improve your SEO. Build up content on your blog, Quora, etc. 
    • Set up as many lead opportunities as possible: after you resolve issues, give people discounts; set up drip campaigns; upsell people who use the product the most; etc.

    Tint's sales tech stack/process:

    • Wufoo forms for inbound sales leads. Wufoo can track sources, which is handy.
    • Gmail for responding to form submissions.
    • Assistant.to for scheduling. Conversions were high because Assistant.to makes it so easy to book a demo.
    • For each inbound request, send a short email that basically says, "Here's a video describing the product, and here are a few times if you'd like to do a call with someone from the team."
    • Pipedrive to track the sales process.
    • DocSend to track docs. (Personal note: doxIQ is another good tool for document tracking.)
    • Boomerang for follow-ups. (Personal note: I like NudgeMail, which is currently free.)
    • Ballpark for invoicing.

    Tricks for scaling up:

    • Sponsor conferences.
    • When asking people to spread the word, offer pre-written email/Twitter/etc. templates.

    Build A Massive Personal Brand-Powered Referral Network by Carolyn Betts

    Speaker background: Carolyn founded Betts Recruiting, which helps companies hire sales and marketing people.

    After you attend a conferences or event: prioritize your follow-ups, make the follow-ups personal, and invite people out socially to form deeper relationships.

    Have a strong online presence:

    • Use professional photos and descriptive titles on sites like LinkedIn. Make sure your profiles are fleshed out.
    • Google your name to see what others see when they search for you. Polish up all of the results that you have control over.
    • Create relevant, valuable content and share it widely.
    • Be active on social sites, blogs, etc. Make connections with readers, and with other writers.
    • Build relationships within your company. This is often undervalued. When people that you have good relationships with move on to other companies, they might be able to get you new clients or help in other ways.
    • Pay it forward. Send referrals, connect people, compliment people to others, etc. The more you do that (in a genuine and honest way), the more people will want to work with you.

    Build a Hyper Targeted Lead List in 12 Minutes by Ilya Lichtenstein

    Speaker background: Ilya is the cofounder/CEO of MixRank, which helps sales teams automate lead prospecting (i.e. find new customers).

    MixRank has grown dramatically over the last year. Ilya talked about how the company did that while spending time on sales, but not on marketing.

    Main theme: you don't need marketing when you're starting out. If you have a product that costs more than a few hundred dollars per month, you don't need marketing, but you will need sales. Blog posts or SEO are helpful, but they're not required. In fact, at a typical company, 80+% of closed deals come from sales, not marketing.

    In the past, qualifying criteria included attributes like job title, company size, location, and firmographics. Today, it's also possible to use tech signals (what technologies does a company use?), social signals (who are they connected to?), fundraising signals, user activity (in freemium/trial product), etc.

    The traditional way to qualify a lead is linear: go through a list of leads and qualify them one-by-one. Today, you can take a more lateral approach: find reasons to say 'no' to most prospects and only focus on the perfect customers. When you're big, you want to add as many customers as possible. When you're small, you don't have enough bandwidth to sell to everyone, so just focus on the customers who are the lowest hanging fruit/the easiest to sell to. You can address other customers as your team grows.

    Think about who your ideal customers are, then figure out what they have in common. Some company attributes to consider: industry, location, funding level, and technologies used. Purchase attributes: seniority, title, and dpeartment. To generate more ideas, first look at common attributes among your own customers, then among your competitors' customers, then among your partners' customers.

    Use Google's advanced search operators to search LinkedIn for new leads. Google shows more results and allows you to do searches that you wouldn't be able to do directly on LinkedIn. Examples:

    Q&A

    Q: How do you find someone's email address?

    A: Use Google/Twitter searches to find people at the same email domain and see what kind of email address convention they use. Then try that convention for other names at the same domain. toofr is a handy tool for this kind of stuff. 

    How to Sell a Product Before You Create It by Josh Isaak

    Speaker background: Josh used to own a brick and mortar business, then went into consulting, then recently moved into SaaS where he has successfully pre-sold products before building them.

    Most businesses fail because they build something that's not wanted. Often, they start building too early and start selling when it's too late. One solution is to sell something before it exists.

    Two key topics covered in the talk: believe you can sell something before you write a single line of code; apply this process to create your first product, or to add new products to your business.

    The process:

    1) Ideas

    Ideas come from problems. Good ideas come from problems of people who have lots of money for solutions.

    So, pick a market that has people with money (dentists, engineers, etc.), and ask those people about their problems. You can just call or email them. You can use a very simple script like: "Hey, I saw your website and really liked __. My name is __ and I'm reaching out because I'm doing research on what problems lawyers are facing. Can you tell me about a challenge that you face? I'd love to hear back, even if it's just once sentence." 

    Email dozens of people a day until you have enough data.

    If you already have a company, ask your current customers about their problems.

    2) Validation

    Is the problem you're thinking of solving universal? Keep emailing and talking to more people until you're sure the answer is 'yes'.

    3) Pre-sell validation

    This is scary =).

    Remember that you have to deliver if pre-selling works!

    How do you presell? 1) build a wireframe, 2) get feedback from potential user, 3) tweak wireframe, 4) go back to user, 5) ask them if they want to join as an early adopter. (Note: users know they're looking at wireframes and not an actual product.)

    For wire-framing, Josh uses Keynotopia. For demos, he uses join.me and Skype.

    After incorporating feedback, invite users to pre-pay to become early adopters. Given them incentives like the ability to be involved in product development and discounted pricing for life. Offer bigger discounts and incentives for longer up-front contracts. For example, you might offer 6-, 12-, and 24-month contracts at 30% off the retail price. If someone signs up for the 12- or 24-month plans, you will give them an additional 2-4 months free. Note that the discount is for life -- the value of great customer advocates is worth more than the lost revenue.

    When you're pre-selling, explain that you want customers to pay so that they can be committed to the product and to giving good feedback, and because pre-selling helps you fund product development. You can also offer a guarantee (e.g. refund within 3 months for any reason.)

    4) Launch!

    First, launch to early adopters and build up case studies. Next, launch to the public. Josh has done this twice now (with clinicmetrics.com and clinicrise.com), and both products were funded with pre-sales.

    Q&A

    Q: Beta customers often want different features, so how do you avoid feature creep?

    A: This works best if you target people in the same industry or with same problem.

    Growing B2B and SaaS Sales Teams by Armando Mann

    Speaker background: Armando build out sales teams at Google, Dropbox, and RelateIQ.

    3 parts to this talk: defining your sales process, measuring success, and maximizing success.

    Part 1: Defining your sales process

    Guiding principles at RelateIQ:

    • Computers are cheap and people are expensive, so...
      • "If a computer can do it, a computer should do it."
      • "If you do something twice, automate it."
      • The team should focus on tough problems that only humans can solve.
    • Use tools and integrations heavily because..
      • Reps do things faster, and that leads to more sales closed. For example, contacting someone within an hour of them singing up can double the close rate!
      • Tools help reps remember things so that their brainpower is freed up for other tasks.
      • Tool usage results in cleaner and better data.
      • Tools improve visibility into what's going on in the company.

    Part 2: Measuring success

    1. Measure key milestones (e.g. what makes someone a lifetime customer?)
    2. Isolate leading indicators for those milestones. What makes them happen?

    Milestones to measure: white paper downloads/engagement (see doxIQ), trial usage, customer satisfaction, renewals, upsells, etc. In general, measure everything. Identify the 3-5 metrics that drive your business. For RelateIQ, it turned out that trials and the brands of existing customers were among the key growth factors.

    Churn is a bad metric to measure (e.g. you might lose 10% of your customers over the course of a year, but be making 2x the revenue with the customers who stayed -- the churn metric looks bad, but your revenue actually grew dramatically). Instead, measure cohort growth over time (ACV by cohort).

    Ways to expand your business:

    1. Add capacity (more companies using your product)
    2. New use cases (e.g. RelateIQ can be used by the sales team, but also the customer success team, the marketing team, etc.)
    3. Adjacent orgs/teams within the same company.

    Random statistic from the talk: 15x close rate (!) for RelateIQ if someone from RelateIQ connects with a self-service client -- even if it's just for a short chat.

    Another lesson specific to RelateIQ is that most people sign up if they've imported data during the trial, so imports receive a lot of focus.

    The recurring theme: figure out what makes people sign up and stay, and then focus on that relentlessly.

    Part 3: Maximizing success

    Most sales orgs are set up like a production line: BD reps get leads, account execs close them, and customer success managers (CSMs) make sure those customers are happy.

    RelateIQ sets people up in "pods" with a biz dev person, a junior and a senior account exec, and a few CSMs. By setting the system up so that the same sets of people keep working together on accounts, accountability goes up. Pods also make it easier to create career progressions. For example, a junior account exec can move to a new pod as their senior account exec.

    Q&A

    Q: Thoughts about pricing?

    A: It's common to give discounts to enterprise, but Armando thinks it's better to go the other way: lower per-seat prices for small customers, higher prices for bigger companies. This could either be done with different license tiers, or with decreasing discounts as company sizes increase.

    Q: Thoughts on free/fremium?

    A: Free makes sense if it's winner take all, or if users don't get immediate value from using product, or if the end user and the decision maker are different. However, there are many situations where not offering a free tier makes a lot of sense -- it depends on the industry and the product.

    How to Write a 3m Page View Article by Nick O'Neill

    Speaker background: Nick organized Sales Conf and used to be a prolific blogger. He grew one content site to 1.4m uniques per month.

    The power law applies to articles. Nick has written 3,500 articles, and 25% of the total page views came from the single most popular article. Given that dynamic, it makes sense to work as hard as possible to produce a hit.

    Miscellaneous tips for content marketing:

    • Backlinko, Videofruit, and Nick's personal site are all great resources for content marketing.
    • Uses Google search suggestions for relevant keywords and ideas. For example, if you're writing an article about Facebook, check which c Google suggests for "facebook a", "facebook b", and so on.
    • Buzzsumo is a great resource to see what's shared by topic/domain. Look for patterns and inspiration.
    • Find a great, heavily-shared article and use the Skyscraper Technique to build something even better.
    • Follow Copyblogger's suggestions for magnetic headlines.
    • Link-building tips: http://backlinko.com/link-building
    • Invite people to do something (share/subscribe/comment/etc) at the top and bottom of every page.

    How Mattermark Sold Their First Million by Danielle Morrill

    Speaker background: Danielle is the cofounder/CEO of Mattermark, which provides private company research, prospecting, and tracking. Prior to Mattermark, she was the Director of Marketing at Twilio.

    Danielle told the story of how Mattermark reached its first million in sales in the course of a year. The talk was more storytelling than bullet points, but highlighted many valuable lessons.

    • When you're selling a brand new product, you need to have very, very high expectations. Because you don't know how high to set your goals, it's easy to settle for sales that seem healthy but are actually suboptimal.
    • Making the first sales hire is scary.
    • The first sales hire needs to be able to run the whole sales process. 
    • Upgrading from a spreadsheet to a CRM for tracking sales is a huge leap forward in effectiveness.
    • Lead velocity is more valuable than lead quality in the early days. If you have a ton of leads, that's a great asset for customer development.
    • Don't just go through questionnaires on your phone calls. You need to understand what your customers are looking for. A good technique is to spend the first 25 minutes of a 30-minute sales call engaging a potential customer, building rapport, and understanding their problems, then spend the last 5 minutes demoing just the product features that they would appreciate. That's a lot more effective than pitching the product in the same, general manner to every prospect.
    • If a potential lead isn't a good fit right now, save your call notes for future reference. The person might be a better fit as your product evolves, and taking good notes will help you close in the future.
    • Good practice: email customers about new features every few weeks. (Personal note: I like this practice and am surprised by how few companies seem to do it.)
    • LeadGenius is effective.
    • When hiring a VP of Sales for a startup, make sure they can function well in a small organization. Sometimes VPs from big orgs look great on paper, but are used to working with expensive tools and well-defined processes and a well-known brand. Those things rarely exist at small companies.

    Sales Hot Seats with Hiten ShahLincoln Murphy, and Steli Efti

    This was a panel Q&A session where attendees would ask the panel how to solve a specific problem.

    Problem: in a previous business, selling was easy because there the product had scarcity built-in. In new SaaS business, that's not the case. How does one create feeling of scarcity -- especially without fiddling with pricing and discounts?

    Suggested solutions:

    • Fiddling with pricing and discounts is not a bad approach. It works.
    • Instead of discounts, offer more product for the same price (e.g. free upgrades/more features).
    • Understand customer priorities and what is urgent for them, then try to connect that to your product.
    • False scarcity can also work. ("Only 3 more 10% coupons available for this month.")
    • Estimate customer's usage and predict the ROI they'd be missing out on. (Example: "you have 12 salespeople and our tool saves 2 hours of work per person per week, which means that at $50/hour, you're losing $1200 for every week you wait.")

    Problem: easy to get people on the phone, but they're rarely the decision makers. Often that's not obvious until well into the call. How do you quickly figure out who the decision maker is without offending the person you're talking to?

    Suggested solutions:

    • If you know who the decision maker would be (e.g. VP of Product), your emails can mention that that's who you're trying to reach.
    • Ask: "in the last 6-12 months, have you championed any tools? What was the process like? What can I do to help with that?" This doesn't hurt someone's ego and makes it easier for them to explain who needs to approve the purchasing decision.
    • If you know what you need to get the deal done, you can give the prospect a checklist so that they know who else to invite to the calls.
    • Treat the prospect as one of several decision makers. "After you approve, is there anyone else who needs to be involved? If you need to present to a team, what concerns might they have that I can help you address?"

    Problem: Product is truly scarce and can only be sold to a certain number of customers. How do you maximize profit/make sure you pick the best customers?

    Suggested solutions:

    • Tell prospects that supply is limited and you want to sell to the people who get the most use out of the product, so you want to determine how much the product would help the prospect.
    • Make prospects go through some hoops in order to demonstrate their level of interest.

    Problem: Sometimes sales cycles take a very long time, especially at big organizations. What can be done to address this?

    Suggested solutions:

    • Ask the potential customer what the process is on their end. For bigger companies with complex processes, ask if you can do some steps in parallel.
    • If you have the flexibility to set a unique price for each customer, ask them what a reasonable time frame would be, then offer a discount that incentivizes them to stick to that time frame. ("You said it should take 3 months? I can give you a 10% discount if it's done within that time frame.") People aren't put off by this and it makes them hustle more.

    Problem: When selling to larger companies, they often want custom work done before committing to a contract. How should that be handled? How should the custom work be priced?

    Suggested solutions:

    • Charge for the pilot.
    • Create scarcity for custom work. You can say something like, "we're very successful with small companies and are now moving to larger deployments. We're doing to work with 10 partners, and we want to work with the companies where this would make the biggest impact on success. If this turns out to be a good fit, you can put down a refundable deposit to be part of our early access program."
      • Note #1: note the use of 'partner' rather than 'customer'.
      • Note #2: note the use of 'early access' (which sounds exclusive) rather than 'beta' (which sounds buggy).

    If you enjoyed these notes, please share them on Twitter.

    Analyzing AngelList Job Postings, Part 1: Basic Stats

    Last month, I used AngelList data to analyze the technologies that startups use. After getting positive feedback on my analysis, I decided to dig into more AngelList data -- this time looking at job postings.  There is a lot of job posting data available via API, and Joshua Slayton from AngelList was kind enough to send me some additional data that is not available via API.  

    This series will consist of 3 separate blog posts:

    Part 1: Basic stats about popular startup locations, job titles, vesting schedules, etc. [Full disclosure: this is the least interesting/actionable post in the series]

    Part 2: Salary and equity benchmarks. By role, company size, etc.

    Part 3: Factors that contribute to successful job postings. [ETA October, assuming my data analysis skills are good enough to come up with useful insights]

    Preliminaries

    I took a snapshot of all jobs on AngelList on August 31st, 2014. There were 12610 jobs posted at the time.

    Please note that this is an analysis of companies found on AngelList, which creates some expected biases. For example, the data will be biased toward earlier stage startups, and away from startups in non-English-speaking countries.

    Job Locations

    Top 20 cities

    City # of open jobs
    San Francisco    2766
    New York City    1676
    Los Angeles    485
    London    444
    Boston    309
    Palo Alto    300
    Bangalore    263
    Washington, DC    230
    Toronto    226
    Chicago    218
    Mountain View    216
    Seattle    201
    Silicon Valley    186
    New Delhi    185
    Austin    161
    Mumbai    153
    Berlin    130
    Vancouver    120
    Santa Monica    117
    Singapore    107

    Next, I removed all cities that had 3 or fewer jobs (about 5% of all jobs) and categorized the remaining cities into geographic areas.

    Jobs in each major geographic area

    Geographic Area # of open jobs
    Silicon Valley    4130
    US (minus Silicon Valley, NY, SoCal)    2282
    New York    1723
    India    839
    Southern California    801
    Continental Europe    649
    Canada    503
    UK    493
    Asia    238
    Central/South America    91
    Australia    78
    Israel    54
    Russia    46

    Notable observations: SF, LA, and NYC make up almost 50% of all jobs on AngelList. Within Silicon Valley, 2/3 of all jobs are in SF and 1/3 are in Mountain View/Palo Alto/etc.

    Openings per company

    # of openings at startup # of startups
    1 opening    3123
    2 openings    1238
    3 openings    689
    4 openings    349
    5 openings    215
    6 openings    115
    7 openings    78
    8 openings    47
    9 openings    30
    10 openings    26
    11 openings    10
    12 openings    5
    13 openings    7
    14 openings    3
    15 openings    1
    17 openings    3
    19 openings    3
    23 openings    1
    25 openings    1

    Notable observations: About 60% of openings are at companies that list <=3 jobs, 35% are at companies that list 4-10 jobs, and the last 5% are at a handful of companies that have 11 or more open positions.

    Job Opening Types

    Type # of open jobs
    Full-time 9740
    Cofounder 1080
    Internship 997
    Contract 936

    Equity Vesting

    Most common vesting periods

    Vesting period # of open jobs
    4 years 11096
    3 years 526
    0 years 355
    2 years 269
    5 years 182

    Most common vesting cliffs

    Vesting Cliff # of open jobs
    1.0 years 11875
    0.0 years 403
    0.5 years 135

    Notable observations: 4 year vesting with a 1 year cliff is by far the most common equity vesting set-up.

    Job Functions

    Role # of open jobs
    Developer 6486
    Marketing 1795
    Designer 1656
    Sales 1601
    Mobile Developer 1434
    Operations 818
    Product Manager 810
    Hardware Engineer 288
    Finance 227
    Office Manager 157
    Human Resources 143
    Attorney 39

    (The counts in the second column add up to a little more than 12610 because some jobs were tagged with multiple roles.)

    In-demand Skills

    Skills mentioned in at least 500 job openings:

    Skill # of open jobs
    Javascript 2372
    Python 1341
    Ruby on Rails 1211
    HTML 1131
    Sales and Marketing 1089
    iOS Development 1069
    jQuery 1016
    CSS 1001
    Java 991
    HTML5 & CSS3 952
    Android 852
    User Experience Design 816
    PHP 777
    Business Development 769
    MySQL 690
    Node.js 644
    Social Media Marketing 604
    MongoDB 509
    Amazon Web Services 503

    Skills mentioned in 250-499 openings: Angular.JS, Objective C, Sales, UI/UX Design, HTML/CSS/PHP/MYSQL, Product Development, Ruby, Mobile Application Design, Mobile Development, User Interface Design, SQL, Backbone.js, Django, PostgreSQL, Linux, Git, APIs, Sales Strategy and Management, Product Marketing, Mobile, Data Analysis, RESTful Services, and Social Media.

    Skills mentioned in 100-249 openings: Big Data, Graphic Design, Hadoop, Adobe Photoshop, Web Design, Web Development, Redis, C++, Mobile Application Development, Photoshop, Product Management, Machine Learning, Software Engineering, SEO/SEM, iOS, REST APIs, Customer Service, noSQL, Twitter Bootstrap, Front-End Development, C, Databases, SaaS, AngularJS, JSON, Business Operations, C#, Growth Hacking, Marketing, Scala, Analytics & Reporting, Adobe Illustrator, Online Marketing, AJAX, Communication Skills, Project Management, E-Commerce, Customer Relationship Management, Full-Stack Web Development, Web Analytics, Sales/Marketing and Strategic Partnerships, Coffeescript, Business Strategy, Account Management, Community Management, Content Marketing, Writing, Backend Development, Amazon EC2, Advertising, .NET, Visual Design, Agile Software Develoment, Software Development, Google Analytics, Adobe Creative Suite, Heroku, and Algorithms.


    If you'd like to be notified when the next part of the series is published, please subscribe to this blog or check back in a week or so.

    The Art of Profitability

    I've read Adrian Slywotzky's The Art of Profitability [affiliate link] three times now, and each time I learn something new. According to its Amazon page, the book "offers 23 business lessons via the tale of a manager's quest to learn the "art of profitability" from David Zhao, a wise master." Each lesson starts with basic principles and uses the Socratic method to build up to a profitable business model. I think current founders, aspiring founders, investors, and employees would find this book valuable and practical. Here are some notes on the profit models that are most relevant to software companies -- although the notes are no substitute for reading the book!


    The main theme: "The path to profitability lies in understanding your customer."

    Assorted Profit Models

    Pyramid Profit

    Key insight: Different segments of customers want different levels of quality/service and have different abilities to pay. Product lines that map well to different customer segments can capture more profit.

    Explanation: A pyramid is a system of three profit tiers:

    • a defensively priced bare-bones product to keep competition out.
    • a standard product for most people.
    • deluxe version to get maximum profit from those with a high willingness to pay.

    Many SaaS businesses follow this profit model: there's a free/cheap tier, a medium priced tier that is appropriate to most customers, and an enterprise tier for customers with deep pockets who want the most features and services. Another example of this profit model would be different car models for a single brand (e.g. Nissan Versa vs Altima/Maxima vs GT-R).

    Multi-Component Profit

    Key insight: Customers can have different price sensitivities for the same item in different contexts. Someone who needs a resume proofread might pay $25, but someone who needs it proofread in the next 24-hours might pay $100.

    Explanation: You can often sell the same thing in different contexts and for different prices. For example, you can sell a bar of soap for $1 at 7-11, a 6-pack $3 at a grocery store, and a 30-pack for $10 at Costco. Similarly, you can rent out a hotel room to individuals, wedding parties, or business conference attendees -- the prices and profitability levels will vary greatly across these use cases, but the room is always the same.

    Switchboard Profit

    Key insight: In contexts where assembling a package of related goods and services takes a lot of effort, customers will pay a premium for pre-assembled packages.

    Explanation: There are situations where a customer prefers to buy a package of several interlocking pieces instead of hunting down each piece separately. The book uses the example of movie studios, which need a director, a script, and a star to make a movie. Studios are willing to pay a large premium to an agent who can deliver a bundle of all three things. Heroku is another example of this model: they offer bundles of software and hardware services, and people pay a premium for not having to deal with installation, integration, and deployment. The more services Heroku supports out of the box, the more it can potentially charge because the cost of configuring, deploying, and monitoring every service individually is too high.

    Blockbuster Profit

    Key insight: In some markets, the path to profit is to produce blockbusters. The movie industry is one such market; the pharmaceutical industry is another one. Because profit in these markets is very hits-driven, companies need to be very comfortable with risk, and they need to understand the components of a blockbuster as much as possible.

    Explanation: In a blockbuster-driven industry (e.g. consumer software, mobile games, etc.), R&D can be a huge money loser if you are doing research in the wrong areas or in an area not worth researching. Focus on understanding blockbusters as much as possible, so that you can focus your R&D spend on the most promising, most profitable areas. It's a shame when someone invests a lot of time and money into developing a product that people don't want.

    Profit-Multiplier Profit

    Key insight: You can improve the effectiveness of R&D by increasing the amount of profit that successful projects produce. If you can turn one successful product line into five then your ROI will be much higher. This is especially true in software businesses where different product lines can share a lot of code and infrastructure.

    Explanation: Many companies can take a single skill or asset and repackage it in many ways. For example, Honda is good at making motors, and they use that skill to sell cars, boats, motorcycles, lawnmowers, etc. Similarly Disney's IP is used for movies, TV, merchandising, music, etc. Unlike the Multi-Component model, which involves selling the same product in different contexts, the Profit-Multiplier model is about creating new products that are repackaged versions of a core asset.

    Specialist Profit

    Key insight: If you understand a problem better than anyone else, you'll be able to create better products, and customers will pay a premium to work with you.

    Explanation: Become a domain expert in a new discipline, then use your expertise to generate profits. Being an expert gives you the knowledge to lowers costs and the reputation to justify higher prices. This model focuses on learning everything about a specific problem. Palantir is a good example of the Specialist approach: by setting themselves up as the foremost experts on data management and analysis, they command higher prices and, as one of my friends said, "they [de facto] have first dibs on any interesting big data problem."

    Installed Base Profit

    Key insight: Customers who already use your products are a great market for upgrades, add-ons, related products, and so on.

    Explanation: One particularly effective business model is to sell products at a low profit margin, then sell add-ons, consumables, upgrades, support plans, and so on at a higher profit margins. Examples of this model include printers (low margin, but ink is high margin), Amazon Kindle (no margin, but ebooks have good margins), and cars (lower margin, but dealer maintenance services are high margin).

    Specialty Product Profit

    Key insight: Specialty products usually earn much higher margins that commodity products (although not for long).

    Explanation: Unique products that serve a small niche can make a ton of money, especially in the absence of competition. However, as the specialty products become more common and more commoditized, their profit margins drop dramatically -- an event that companies need to plan for. A lot of recent SaaS services, which are becoming more and more specialized, are good examples of this model. There are now CRM programs designed for dentists, cloud-based auditing tools for CPAs, and the like.

    Local Leadership Profit

    Key insight: Having a near-monopoly is one geographic area can be more profitable than owning a small piece of the market across many locales.

    Explanation: This is the playbook for a lot of on-demand startups: Homejoy, Uber, Move Loot, Spoon Rocket, etc. If you're competing with a lot of companies for the same location, profits plummet (or go negative). On the other hand, if you have a monopoly on a location, you can enjoy high profit margins while your competitors struggle to get even a small foothold.

    Relative Market Share Profit

    Key insight: The higher your market share, the more advantages you have in terms of cost structure, distribution, marketing cost per unit, R&D cost per unit, and so on.

    Explanation: The biggest player in the market can spread their fixed costs across many more units, which provides the flexibility to decrease prices or increase advertising spend or take other actions that make it even harder for others to compete. Companies that understand the importance of relative market share do whatever it takes to become #1. Think Uber vs. Lyft as a great recent example.

    Another advantage of higher market share is that you can accumulate more data, which increases competitive barriers. For companies like Yelp (business review data), LinkedIn (professional network data), and Google (search query data), the ability to learn from more data makes it next to impossible for someone to compete, even if they have a better fundamental product. For example, someone who makes a better version of LinkedIn will still struggle to get anywhere because LinkedIn has so much more data about what kind of job recommendations people respond to, how to find colleagues that someone might know, how people in different companies are connected, and so on.

    [Data network effects are a key reason for my fund, Susa Ventures, having a strong focus on companies accumulating valuable datasets.]



    I think that the best way to use The Art of Profitability [affiliate link] is to try applying different models to your own business and see if any of them could fit. Is there a core asset you can repackage into different products? Do customers have different price sensitivities for your product in different markets? Can you sell your product as part of a pre-assembled bundle to save your customers from integration headaches? And so on.

    Which Technologies Do Startups Use? An Exploration of AngelList Data.

    There's a lot of hype surrounding new programming languages, databases, and the like. I've always been curious about which technologies are actually in use, and whether great startups use different technologies than not-so-great startups.

    Fortunately, AngelList offers some self-reported data about technology usage. For example, you can see that Robinhood uses Python, Django, and iOS while Secret uses Java, Go, Python, JavaScript, HTML5, CSS, iOS, and Android.

    Additionally, AngelList calculates a Signal score for each startup. While it's not 100% clear what this represents, it seems to be some combination of company quality and popularity. For example, these are all of the startups in the Transportation sector, sorted by their Signal scores.

    In the interest of openness, there are lots of caveats for this dataset:

    • It's not clear if AngelList Signal is actually correlated with company quality (although it seems to be).
    • Many companies don't report the technologies that they use.
    • The lists of technologies that are self-reported are not necessarily exhaustive.
    • etc.

    Limitations aside, I calculated the number of startups with low, medium, and high Signal scores using each of ~75 different technologies, and this post summarizes the results. Whenever I refer to okay/good/great companies, the intended interpretation is companies with low/medium/high AngelList Signal scores.

    (Note: you can click on each chart to see a higher resolution version.)

    Interpreting the Charts

    In each chart, blue represents 'okay' startups, red represents 'good' startups, and orange represents 'great' startups. Within each color, the bars show relative frequencies of technology mentions. For example, let's say we're looking at technologies A and B. If 'okay' companies use B 3x as often as A, 'good' companies use A and B equally often, and 'great' companies use A twice as often as B, then the chart would look like this:

    (Note: the ratios of the lengths of blue/orange/red bars are 1:3, 1:1, and 2:1.)

    Programming Languages

    Observations:

    • For great companies, the most popular languages are Javascript, Ruby, Python, and Java.
    • For okay companies, the most popular languages are Javascript, Ruby, PHP, and Java.
    • The likelihood that PHP is being used is strongly anti-correlated with company quality.
    • The better the company, the more likely it is to be using modern and/or functional programming languages (i.e. Go, Scala, Haskell, Erlang, Clojure).

    Front-end Tech

    This chart shows a combination of web frameworks (Rails, Django), Javascript Libraries, and HTML/CSS.

    Observations:

    • Ruby on Rails is super popular.
    • HTML5 is dominating HTML.
    • CSS is still dominating CSS3.
    • The better the company, the less likely it is to use Bootstrap. 

    Databases/Storage/Caching

    Observations:

    • MySQL, Mongo, and Postgres dominate the database side.
    • Redis is much more popular than memcached.
    • The better the company, the less likely it is to build on top of Microsoft's products (SQL Server).

    Mobile

    Observations:

    • Developing for iOS is slightly more popular than developing for Android
    • The gap widens as company quality increases.
    • Windows Mobile (which is not present in the chart) is 30x-50x less popular than iOS and Android among good/great companies.

    Infrastructure/Hosting

    Observations:
    • AWS and Heroku dominate.
    • The better the company, the more likely it is to use IaaS (e.g. AWS) instead of PaaS (e.g. Parse)
    • The better the company, the less likely it is to build on top of Microsoft's products (Azure).

    DevOps Tools


    (Note: the sample size here was small because DevOps tools were rarely mentioned on AngelList profiles)

    Search

    Observations:

    • Elasticsearch dominates this category.

    API Integrations

    (Note: the sample size here was small because APIs were rarely mentioned on AngelList profiles -- especially for good/great companies.)

    Advanced Tech

    Observations:

    • I was surprised that there's no clear correlation between quality of company and usage of sophisticated technologies like machine learning or computer vision.

    Big Data Software

    (Note: small sample size)


    If you liked this post, please consider sharing it or subscribing to future posts.

    Beneficiary/User/Purchaser Alignment

    For founders, Product/Market Fit (PMF) is the holy grail**. Achieving PMF means you have developed a product that satisfies a strong market need. As Marc Andreessen once said, "the only thing that matters is getting to product/market fit."

    While finding PMF is critical to a startup, there is one nuance that is often overlooked, which that even a great product that fulfills someone's needs can struggle if it fails to align with organizational incentives. Specifically, a typical product involves several major parties, including the user, the purchaser, and the beneficiary. The user is the person using the product day-to-day, the purchaser is the person who pays for the product, and the beneficiary is the person who benefits from product usage. When selling to large, complex organizations, it turns out that the user, the purchaser, and the beneficiary are rarely the same person.

    Here are a few examples:

    • For CRM software, the main user is the sales rep, who needs to enter data into the system after each sales interaction. The main beneficiary is often the VP of Sales who wants better data about the sales pipeline. Sometimes the purchaser is the VP of Sales, but sometimes it's another department, like IT.
    • For an app that combines personal tracking (e.g. from a Fitbit) with Electronic Health Records, the main users are the patient and the doctor; the main beneficiaries are the patient, the hospital, and the insurance company; and the payers are likely the insurance company and/or the hospital.
    • For data analysis tools geared toward less technical business people, the main beneficiary is the business person but the main user is typically the data analyst or the IT person who sets up the tool. The purchaser could be any of those three.

    The difficulty in each of these examples is that solving one party's problems might not matter if another party is footing the bill or forced to do all of the work. For example, CRM tools that are designed for VPs of Sales often have adoption problems because salespeople have little incentive to use them. On the other hand, CRM tools that are designed to make salespeople more productive are hard to sell because VPs of Sales are the ones who have to pay for them, but they don't get many direct benefits.

    As you develop your product and find yourself reaching product/market fit, think about every part of an organization that your product will touch. If your users are not your purchasers, or your purchasers are not your beneficiaries, tweak your product to align with everyone's incentives. For example, if users are salespeople, beneficiaries are sales managers, and purchasers are IT, then you could focus on day-to-day productivity enhancements for salespeople, great reporting and analytics for sales managers, and solid security and 1-click deployment for IT. Don't just focus on reporting for sales managers, because even though you'll be solving their problems in theory, you'll either have a hard time getting salesperson engagement, or you'll get pushback from IT because they don't want to manage yet another 3rd party service.


    ** In fact, if you search for "startup holy grail" on Google, three of the first four links are about product/market fit.


    If you liked this post, please consider sharing it or subscribing to future posts.

    Buzzwords That Backfire

    Startup pitches are often brimming with buzzwords:

    • "We're revolutionizing HCM with Big Data and Machine Learning!"
    • "We have a team of NLP ninjas and Ruby rockstars!"
    • "We're already at $1mm ARR!"

    Trying to get people excited about your startup is good, but hype and buzzwords can backfire if they're misleading. As an engineer turned VC, I see a lot of pitches whose audience is typically business folks, and where technical buzzword usage is liberal and sometimes manipulative.**

    Here's a typical conversation:

    Founder: "Our elite engineers are building cutting edge machine learning algorithms to predict stock market prices."

    Me: "Can you elaborate on machine learning a little bit?"

    Founder (who is used to talking with investors coming from a business/finance background): "I mean, you know... Machine Learning, man. It's like all the rage these days."

    Me: "Yes, I know. I'm an engineer. Please tell me about one or two of the heuristics that you use."

    Founder: "Oh. Well, actually, we use some 3rd party software that does the machine learning stuff for us, so I'm not sure what that's doing under the hood."

    So much for building cutting edge algorithms. At this point in the conversation, I make a mental note to pass on the investment. There's nothing wrong with using 3rd party software to do some of the heavy lifting, but there's a big problem with trying to mislead investors. As a VC, I immediately assume I'm looking at an iceberg. That is, if I've uncovered a few false statements, then there are likely many more beneath the surface. Even if the rest of the pitch is good, it's too hard to build a strong founder-investor relationship when there's no trust. 

    If you're pitching investors, please remember that while it's fine to reference hot trends and buzzwords in your pitch, make sure you do so honestly. Don't claim to do things that you don't do. Hiding behind trendy terms is risky because: 1) eventually you encounter someone who can see that you're misrepresenting yourself and 2) that person, when asked, will share their findings with other investors.


    ** To be clear, this is not a knock against non-technical VCs. Someone like me is as prone to being misled by business or financial claims as they are to being misled by technical claims.


    If you liked this post, please consider sharing it or subscribing to future posts.

    Benefit Providers for Startups

    My fund, Susa Ventures, recently conducted its first annual founder survey. We asked each of the founders we work with about a dozen questions about everything ranging from recruiting to benefit providers to company culture. The results were fascinating, and I'd like to share some of the data about benefit providers. If you're starting a company and are not sure where to turn for services like payroll or health insurance, then I hope this blog post will provide some good starting points.

    Reporting methodology

    Satisfaction scores are on a 1 to 5 scale, with 1 being "very dissatisfied", 3 being "neutral", and 5 being "very satisfied."

    All costs are reported as per employee per month (/e/m).

    Caveat

    The sample size is not huge (17 founders replied, and some founders skipped a few questions). The following recommendations are often based on 3-5 founders all recommending the same service. Given the sample size, please take all recommendations with a large grain of salt.

    Payroll

    Typical cost: $5-$10/e/m.

    Best provider: ZenPayroll. About half of the respondents used ZenPayroll, and their average satisfaction score was 4.7. Everyone who used ZenPayroll was very happy with it.

    Health Insurance

    Typical cost: $200-$600/e/m.

    Best provider: No clear winner given fragmentation and variance between health plans. One interesting observation was that the more companies paid, the happier they were. Companies paying $200-$300/e/m typically had satisfaction scores of 3 or lower, while companies paying $400+/e/m had satisfaction scores of 4 or higher.

    The two most popular providers were Zenefits (5 votes) and Anthem (3 votes). Both were in the $350-$425/e/m range and both had average satisfaction scores of 3.7.

    Dental Insurance

    Typical cost: $17-$60/e/m.

    Best providers: 3 founders were satisfied with Delta Dental (prices were $20-$60/e/m). 2 founders were very satisfied with Zenefits ($17/employee/month).

    Legal

    Typical cost: highly variable.

    Best providers: Gunderson Dettmer was unanimously liked by the 3 founders who used it (4.7 satisfaction score). Cooley and Wilson Sonsini each had 2-3 votes and satisfaction scores of 4.0.

    Accounting

    Typical cost: $500 per company per month.

    Best provider: inDinero had 4 votes and an average satisfaction score of 4.0.

    Worst provider: Doing accounting by yourself had an average satisfaction score of 2.0.