How to Evaluate a Startup (for Success)
This post is part of a series titled, “How to Select Your Next Job”. I’d love to hear your feedback in the comments section below!
If you’re a Software Engineer, Product Manager, or Data Scientist without a tech background, it can often be difficult to evaluate startups. There are hundreds of thousands out there — how do you know which one is going to succeed or fail? Which ones should you bother interviewing at? What exacerbates this challenge is that as an employee, you have a portfolio of one. As opposed to being an investor where you can spread your money among many different companies — when you’re employed, you’re exchanging your time for equity in only one startup, where you’ll likely spend at least 2–4 years working. And unless you have a financial background (i.e. you worked in investment banking, venture capital or private equity), you really aren’t likely equipped with the tools for deciding which company to pursue.
Luckily, I give laypeople an easy rule of thumb for evaluating the financial success of a startup they’re considering.
For Early-Stage, look at Team
For Mid-Stage, look at Market
For Late-Stage, look at Growth
Note — this is not necessarily investment advice. If you’re a finance person who can do precise modeling, by all means do that. This advice is geared towards tech employees who don’t come from finance.
In order for this rule of thumb to make sense, we have to first define what each of these is. I define things in the following way (don’t hate on me if you’ve heard a different definition):
- An early-stage startup is one that hasn’t found product-market fit.
- A mid-stage startup is one that has achieved product-market fit, and is growing the adoption of its product within that initial market.
- A late-stage startup is one that has reached saturation in its initial market, and is growing the overall size of the market and/or expanding to new markets.
Note — the way that I’ve defined these terms, has nothing to do with things like funding, revenue, or headcount (per se). A 10-person startup could be late stage if it’s cornered a niche market and is expanding to other markets. A 1000-person company with billions in funding could still be early-stage if it hasn’t found product-market fit. Let’s run through a few examples:
- Sarah and John are two entrepreneurs working on a social application to help doctors keep track of patient data. Primary care physicians and specialists can take notes on their patients and share them with each other in a HIPAA compliant way. This business requires a long sales cycle and a large army of salespeople, engineers, and compliance operations specialists. Because Sarah and John have had successful exits before and strong domain expertise, they have no problem raising money; in fact they’ve already raised up to $50M and have a team of 40 people. They’ve already built a prototype that they’ve shared with some of the doctors in their family, but to get more adoption, they’re working on a second iteration of the product. This (per my definition) is an example of an early-stage startup. Despite money raised and headcount, they haven’t found product-market fit yet.
- Coinbase is a startup that allows people to purchase cryptocurrencies. With 11.9 million active users who use the app to purchase Bitcoin, they are quietly taking over the reserve digital currency wallet market. This is an example of a mid-stage startup. It’s found product-market fit (a product that fits its market), and is growing its share within that market.
- Uber is a startup that allows users who download its app to hail a vehicle to take them somewhere. It operates in the private, for-hire, on-demand transportation market (aka the Taxi market), which prior to Uber’s existence was something like $100B globally on a bookings-basis. Based on what’s been reported, the company has already occupied a sizable chunk of the market, and via disruption, is actually growing the overall size of the market. It’s also expanding to adjacent markets (food delivery). Based on my definition, Uber is a late-stage startup.
At each stage, the most important factor is the likelihood the startup will get to the next stage.
Early-Stage Startups: All About the Team
As Marc Andreessen (some cool hipster VC) writes,
The only thing that matters is getting to product/market fit. When you don’t have it, it’s all that matters. When you have it, nothing else matters.
Because an early-stage startup hasn’t achieved product/market-fit (PMF), when you (the engineer, pm or data scientist) are considering it, the only thing that matters is the likelihood that the company can achieve PMF. Nothing else is material in terms of evaluating the potential success of the company. Good companies at this stage are constantly shifting their product or market around, thus making an evaluation of other factors like product, market, or growth irrelevant. The only thing you should be asking yourself is “To what degree do I believe that this team can find product-market fit quickly?”. During this phase, ideas are cheap. Being dazzled by a cool product prototype is pointless — there’s no guarantee that it’d satisfy a market.
Note — when I say that team is the thing that matters, it doesn’t mean to find the most pedigreed team, most technically competent team, or most famous team. The quality of the team is determined by how quickly the team can find product-market-fit. When your initial stab at PMF doesn’t work, you have to either modify the product, or try a different market.
Consider the following example. Sarah and John are entrepreneurial superstars, with multiple exits in the past. To capture the doctor market for their startup, they’ve hired an all-star team of ex-Pharmaceutical reps who are the top ex-Merck/ex-Pfizer salespeople. They’ve also hired Staff-level ex-Google and ex-Facebook engineers to build a centralized patient information system. These engineers have contributed to open-source systems libraries like Cassandra and are capable of architecting high quality, scalable, bug-free systems that are extremely secure, which is important for housing patient data. However, because their startup hasn’t achieved product/market-fit, none of this matters. Sarah and John find that selling to doctors isn’t the right approach and that the real market should be consumers who store their own patient data on an iPhone app.
Evaluating the team at an early-stage startup, isn’t necessarily about finding genius or pedigree, or serial entrepreneurs¹. Ask yourself, “given the space this company is operating in, how confident am I in this team’s ability to find product/market fit quickly?”
The final reason team is so important at the early-stage is that the majority of the intrinsic value of the startup comes from the team. The price that someone would pay for the team in an acquihire can serve as a floor or “safety net” to what the company could fetch in an exit, provided it doesn’t achieve any form of traction.
Questions to Ask:
- Is the team well-equipped to handle change rapidly? If in finding PMF, they had to either change the product or the market, how easily could they do so? The best teams are able to change both on a dime.
- How quickly can the team move? Is there some attribute about this team that gives it an unfair advantage to other pre-PMF teams trying to operate in the same space?
- Does this team have some specific skill set(mobile in ’10, VR/AR in ’14, self-driving automation in ’16, etc.) that an acquirer would pay a premium for?
Questions that are Irrelevant:
- What is the market opportunity for the product being built? (PMF not reached yet, so market is not defined to make a call.)
 That’s not to say that a team of geniuses or domain experts or serial entrepreneurs is necessarily a bad thing. If the startup is in a space that requires you to be a genius to reach product/market fit quickly, then it might actually be a good thing.
Mid-Stage Startups: All About the Market
So you’re considering a startup that’s already found PMF. This means the company’s already shipped a minimally viable product that satisfies at least some portion of the market. What matters now is evaluating how good the market really is (and the company’s strategy for capturing it). There are many different factors to consider (here are about 10 of them). However, if you’re an engineer like me (and not a business/finance person), it can be difficult to evaluate the startup against all of these, so I generally scope it down to just four factors:
- Size of the market
- Is the market growing?
- Unit economics/metrics of the market
- Capturability and Defensibility
What is the size of the market?
This seems like a no-brainer. Entrepreneurs or recruiters often like to brag about the size of the market their startup is disrupting (“The used car sale market is $2 Trillion!!” or “The enterprise payroll market is $400B!”).
The first point here, is that what really matters is not precisely how big the market is, but simply whether the market is big enough. Once the market is of a large enough size, diminishing returns catch up pretty fast (one company that boasts a trillion-dollar market is really not that much better than another one that boasts a hundred-billion-dollar one). The second point is that companies will define whatever addressable market they think makes them look better. It’s up to you to decide how much of that market their product is actually capable of capturing. For example, if you receive an offer from a startup that helps you mine bitcoin, they might claim their addressable market should be pegged at the entire size of all reserve currency + gold (in the trillions), and you might argue the addressable market is just that of the major cryptocurrencies bitcoin+ether (tens of billions). My general guidance on this is, for mid-stage startups, generally go with the smaller one².
How fast is the market growing?
This is substantially more important than overall size of the market. Choose a small-but-fast-growing market over a large-but-stagnant market every single time.
The entire mobile advertising market in 2010 was $600m (tiny). At the time if you were just going off of total market size, you wouldn’t have likely concluded that two startups (Facebook and Twitter) would be raking in billions each in mobile ad revenue in a quarter, within 6 years. Mobile advertising, from 2010–2015 grew more than 33x, more than doubling each year. A rising tide lifts all ships is a truism, even with poor execution, companies performed well in this space.
Markets that are large but stagnant generally have an incumbent. For the company to grow its share, it requires competing and taking market share away from the incumbent. Markets that are naturally growing present opportunities for companies to grow without taking share away from incumbents. The former, at minimum, requires a product that is either differentiated, cheaper, or better than the incumbents’ in some facet (or more than one facet). The latter does not. For mid-stage startups with PMF, how fast the market is growing is perhaps the most important factor to consider. 
What are the Unit Metrics like for the market?
There are two specific unit metrics that are important when considering the market that your mid-stage startup is in: LTV and COA.
- Lifetime Value: What is the lifetime value for users in said market?
- Cost of Acquisition: What is the cost of acquisition for customers in said market?
- How much revenue will each customer net, on average to the business throughout the time that they are a customer of the business?
- In order to compute this, you need to do some cohort analysis and understand the average churn for a customer.
- You use the average revenue each customer delivers per time period divided by the average churn rate for that time period.
- Note: there are likely more advanced models for this, but for your analysis this is all you need to know.
- For example: Let’s take the market of all potential ride-sharing customers in Noe Valley in San Francisco. If the average ride-sharing customer is willing to take $50 worth of trips (in terms of gross profit), and without any promotions or marketing has a 20% chance of dropping off of a ride-sharing platform, then the average lifetime value in this market would be $250.
- Note: For any given market (and thus business) there are fixed and variable costs. You should also evaluate: 1) in all likelihood you should incorporate variable costs into your lifetime value computation and 2) to what degree even variable costs might decrease due to things like economies of scale.
- Note: Generally speaking, Lifetime value generally applies to customers within the same market, so businesses competing for the same set of customers will have the same lifetime value (so Lyft’s and Uber’s customers are likely to have the same lifetime value). This assumption breaks down if one business is able to generate some “sticky” advantage through either branding (e.g. Apple loyalists), market power (e.g. Amazon), network effects (e.g. Facebook’s social graph).
Cost of Acquisition:
- How much does it generally cost (on average) to acquire customers?
- Generally for any given market there are multiple ways to acquire customers. When evaluating the startup’s market, you should try to do your diligence across a wide variety of current and possible acquisition opportunities.
- This can be things like: CPA for Google/Facebook ads, Sales salaries and bonuses, Marketing expenses.
- To compute this with precision, model the customer acquisition funnel and measure the cost and conversion at each point in the funnel.
- For example: The business has discovered the most cost effective way to acquire new customers is through Google Adwords. The average Cost per Click charged for an ad unit is $0.20, and once the customer lands on the website, there’s a 2% chance the user will go through the signup flow for a free trial. Finally, 20% of users who complete the free trial convert into a paying user. The average cost of acquisition for a paying customer is therefore $0.20 x 50 x 5 = $50.
What is the Capturability and Defensibility of the market?
Given the company has found product-market fit, it’s also important to understand two key additional components of the market: “capturability” and “defensibility”.
- Do the dynamics of the market make it easy in particular for your product to capture (more so than other competing products)?
- Once your product captures the market, how easily can the product “defend” against other products that are also trying to capture the same market?
- Usually capturability and defensibility of the market go hand-in-hand. This means that in markets where your product can easily capture new customers, other products can also more easily turn around and take those customers away from you.
- Think through the nature of the product-market fit that the startup’s product has delivered. Has it been achieved through easy user experience and setup? Is it because it’s priced very low? How easy would it be for a competing product to also replicate the same level of success?
- Think about if the barriers to entry work in the product’s favor or against it.
Barriers to Entry
There are many different types of barriers to entry.
- “First-Mover” (low). Customers prefer one product simply because it was there earlier. By being there earlier, you gain advantage and don’t incur switching costs to get your product in front of customers. For example, if you’re a content or media company, being there earlier means that you have content that others don’t have. The reason this is a weak barrier to entry is that over time it will erode unless cultivated — meaning you have to continuously spend $$ to continue to stay ahead of others.
- Economies of Scale (medium). The second class of barriers to entry surround economies of scale. In this scenario, being massive gives your cost and value structure decisive advantages. Bigger = Better. Amazon, Google, Uber are all good examples of this. One common under-appreciated way in which economies of scale manifest itself is in the use of data — it’s common that the more data a product has, the better that product can end up being (for ex: Search and Newsfeed fall in this category).
- Network Effects (high). This is the highest type of barrier to entry — each person on a network adds to the aggregate value of the network in a super-linear way. Ex: Two-sided marketplaces, Telephone, Email, Microsoft Office, Facebook.
Can your mid-stage startup, that’s developed product-market fit leverage barriers of entry to its advantage? When Google was starting out, it operated in a world where the barriers to entry were low — the only ostensible advantages other technologies (physical libraries, Yahoo’s manual directory of the web) had were first-move advantages (e.g. they had built up indexing of content). Google over time turned the market into one in which economies of scale would dominate (the more searches operated on Google’s platform, the better data it had to improve its search functionality).
You might ask…
Why shouldn’t you look at Team when evaluating startups at this Generally speaking, team does make a difference at every single stage (CEOs do after all get paid the big bucks at all stages of the startup). However, once a company has reached product-market fit, the market itself is usually what dominates the outcomes.
This point is particularly important as Marc Andreessen and Andy Rachleff have said multiple times — if you are an awesome founder in a bad market, you’re dead, but if you’re a below average founder in an amazing market, you’ll still likely build a good company.
If you’re an engineer trying to pick which startup to join, at each stage, you’re effectively trying to compute the probability that startup makes it to the next stage successfully, and once a company has found product-market fit, the overly dominating factor is the characteristics of the market.
Why shouldn’t you look at Growth when evaluating Mid-Stage Startups? Good question. The easy answer is that growth can be faked at this stage. There’s no scarcity of techniques people can use to juice growth numbers. Companies can pay for app installs, Facebook and Twitter ads, offer promotions, etc. However once a startup hits a certain size it becomes unrealistic and cost prohibitive to do so.
Questions to Ask
If you’re an employee trying to evaluate whether to join a mid-stage startup, here are the questions to ask yourself:
- Given the company has product-market fit — what is the current size of the market that the existing product is a good fit for (e.g. perhaps the product works for one subset of the market but will need to be improved for the rest of the market)?
- How quickly is the market growing over time?
- What are the unit economics/metrics of the market? Is the Lifetime value significantly (multiples) higher than what I anticipate the cost of acquisition to be? Do people churn frequently in a way that would cause LTV to be low?
- Is the product that has developed product-market fit well-positioned to make use of the barriers to entry that will develop?
 The counterexamples to these are companies that fundamentally grow the market they operate in (e.g. Uber has grown the taxi market, Airbnb the hospitality market, i.e. people are taking Uber today who never took a taxi before). However, since we’re talking about mid-stage companies, for now I’m purely evaluating them on the likelihood they become a late-stage company and ignoring what happens once they saturate their initial market. At least for these two deca-corns, the open question of whether the market was big enough at the time is a resounding yes.
Late Stage Startups: All about the Growth
I define a late-stage startup as one that has already reached a significant share of its natural market, and is now looking to either: 1) start to actually grow the natural size of its intrinsic natural market or 2) fundamentally play in additional markets. Where to draw the cut-off line for when a startup pivots from a mid-stage to a late-stage startup is often blurry/unclear. Here are some examples for delineation:
- For Yahoo, they pivoted to becoming late-stage around 1996/1997, when the company moved their original goal of organizing the internet around hierarchy, and started investing in the overall broader growth of the internet (e.g. acquiring geocities, RocketMail, and building Yahoo! Groups and Messenger).
- For Uber, the company started morphing into a late-stage startup around 2014/2015, with the launch and investment of UberFresh (later Eats), investments in Self-driving, and expansion of UberRush.
- For Google, I use 2004/2005 as the delineation, as this coincided with several key moves (development of Gmail, acquisition of Youtube) that signified the business was expanding beyond its core market.
So why do I encourage employees looking to join a late-stage startup to focus primarily on growth (rather than market or team)? A few reasons:
- Startups that reach late-stage have already demonstrated mastery over their given market, and have had a team in-place that knew how to grow a product to market.
- Many of the subsequent expansions of the business beyond the initial market require well-capitalization and success within the given market that is measured through its growth.
- Once the company reaches late-stage, then at least a sizable amount of success of the company is already guaranteed. The employee’s prospects therefore are primarily tied to how quickly that company is growing.
- Things like “team” and “market” become less meaningful, as the company has a very large “team”, and a diversified set of markets it is pursuing. Therefore evaluating them has less of an outsized impact on your success at the company.
- Late-stage companies generally can no longer “fake” growth through unscalable means (paying for the growth). What you’re looking for is prolonged, sustainable, proven growth at large scales. Early-stage or Mid-stage companies can often see growth spurts only to drop out (think Yo or Yik Yak). But such growth surges that occur at late-stage are more likely to be legitimate.
So what are some ways to evaluate whether the late-stage company you’re joining is growing quickly or slowly?
- LinkedIn Premium will tell you if the late-stage company is hiring aggressively and growing Headcount at a fast pace.
- Google Trends will tell you if there is very large growth in organic search interest.
- Premium analysis tools like AppAnnie and Evercore ISI will tell you download, usage, application metrics.
There you have it! You don’t have to be a valuation expert, have an MBA, or be ex-Finance to evaluate the startup. Just follow these three rules: For early-stage startups evaluate team, for mid-stage startups evaluate market, for late-stage startups evaluate growth.