top of page
Search
Writer's pictureChris Yeh

The Blitzscaling Scorecard: Identifying Future Market Winners

Updated: Oct 11

All the members of the Blitzscaling Ventures community know that the Blitzscaling Scorecard is a key part of what makes Blitzscaling Ventures unique.  It’s a systematic approach to assessing potential investments that is designed to allow us to quickly identify which startups are potential blitzscalers, and thus have the best chance to win a valuable winner-take-most market and deliver great returns to our investors.


The Blitzscaling Scorecard Uses seven key elements of blitzscaling to calculate an overall 1-100 score:


  • The winner-take-most market dynamic from network effects and land grabs (29 points)

  • Scalable distribution: viral effects, or tapping existing distribution networks (29 points)

  • Product-market fit (14 points)

  • Market size (11 points)

  • Gross margin (7 points)

  • Organizational scalability (5 points)

  • Operational scalability (5 points)


Companies that score 80 points or above out of 100 on the Blitzscaling Scorecard have the best chance of successfully blitzscaling and winning a valuable winner-take-most market.


We use the Blitzscaling Scorecard to assess hundreds of startups per month to quickly filter our deal flow down to the approximately 5% of startups that fit our model.  The scorecard is even more important in the Age of AI; as we witness a Cambrian explosion of new technologies and business models, it is essential that we invest in startups that can build a sustainable competitive advantage that leads to enduring market leadership, rather than being overtaken by the next month’s newer and shinier objects.


But very few people have known the origin story of the scorecard, and how it evolved over time.  Until now.


This essay explains how the Blitzscaling Scorecard came to be, how it evolved over time based on the data, and how we use it today.  While I will necessarily obfuscate some of the details to protect our secret sauce, you will have a much better understanding of how Blitzscaling Ventures works after reading this piece.


The roots of the Blitzscaling Scorecard go back to when Reid and I were writing the Blitzscaling book. We knew we needed a way to explain to our readers when to blitzscale, and when to pursue a slower-growth strategy. After we taught the Blitzscaling class at Stanford in the Fall of 2015, we assembled a list of key growth factors which were necessary to successfully blitzscale, and key growth limiters that had to be overcome to do the same. The key growth factors were a large potential market, high gross margins, network effects, and scalable distribution. The key growth limiters were a lack of product market fit, and an inability to scale the organization and operations.


When Jeff Abbott, Scott Johnson, and I set out to create Blitzscaling Ventures, we started with a simple model of assessing each deal by scoring it on a 1 to 10 scale for each of the seven factors of Blitzscaling, but we found that as we applied this simple model to assess potential investments, it didn’t produce the results we wanted.  Certain companies that were clearly successful blitzscalers, like Airbnb and Uber, saw their scores dinged because of issues like operational complexity.


We realized that the growth factors and growth limiters were not created equal.


Our next step was to apply one of my favorite techniques, a weighted decision matrix.  It seemed clear that the most important elements were network effects and scalable distribution, so we weighted these more heavily than the others (eventually, we changed the element to “network effects/land grab” because there can be certain forms of strong switching costs that are due to factors other than network effects).  But as we tweaked and tested the matrix, we now had a different problem.  Companies that scored pretty well but not outstanding on those key factors were getting through the filter, and we seemed to be screening out other companies that scored very well in those categories, but had flaws in less critical areas like operational scaling.


It was then that Scott came up with a simple but brilliant solution.  He added an exponential factor to our formula.  Instead of a simple linear 1 to 10 score, the exponential factor (which is now one of our trade secrets) makes outstanding scores disproportionately valuable.  This makes for a better filter, since it is generally easier to fix a low score than it is to push a pretty good score into the outstanding range.  We experimented with a number of different weighting factors, and tested them against five years of company data before settling on a final formula that produces a single Blitzscalability score on a 100-point scale.  Startups that score 80 or above meet the threshold for blitzscalability.


This is still the Blitzscaling Scorecard that we, Blitzscaling Ventures, use as a critical-path element of our investment process.  Here’s how we do it.


  1. The General Partners of the fund gather for a synchronous meeting.  Generally, we use Zoom for this, though we try to take advantage of running a Scorecard session in person on the several times per year we are gathered together.  For Blitzscaling Ventures I, this means that Scott, Jeff, and I score the companies, and for the Blitzscaling AI Fund, the three of us and Jeremiah Owyang score the companies, with Jeremiah taking the lead as the lead GP for that fund.

  2. All of us are looking at the same Google Sheet, into which our spreadsheet wizard Scott Johnson has built the Blitzscalability formula.  This is the tool we use for our scoring, and no, we can’t share it with you (though you can find a simplified version of it at the Blitzscalabilty Grader website.

  3. For each startup, we grade one element at a time, starting with the winner-take-most market dynamic, followed by distribution.  This is a group process.  Rather than each partner grading the elements individually and feeding scores into an overall average, we believe that there is benefit to debating the scores, sharing thought processes and examples, and arriving at a consensus score.  This doesn’t mean that we all agree, but we do agree on what score to use for the purposes of the exercise.  We have a formal rubric to guide the scoring, but there is necessarily an element of human judgment involved.  (It will come as a surprise to no one that we have experimented with using AI for this, but we still aren’t happy enough with the results yet!)

  4. Once we enter in all the raw scores, the scorecard calculates the exponential factor and weights the results to a 100 point scale.  The two most important factors, the winner-take-most market dynamic and the scalable distribution together can account for up to 58 points.  The third most important factor is product-market fit, which can account for up to 14 points.  After that comes market size (11 points), gross margin (7 points), and finally organizational and operational scalability (5 points each).

  5. One of the ways we make the process efficient is that we know that there are certain patterns to scoring.  For example, a pure software product will generally score 10/10 on organizational and operational scalability, since the simplicity of bits-based businesses almost always allows them to operate in a people- and capital-efficient manner.  We also know that our formula is such that the minimum combination of winner-take-most and distribution scores that can still reach the magical 80-point threshold is 9/8 or 10/7.  In other words, a company that scores an 8 for winner-take-most and an 8 for distribution cannot possibly reach our 80-point threshold, and thus is eliminated from consideration.  This requires intellectual honesty, because it’s easy to manipulate the system by scoring a deal you love high enough to pass.  That is why the Blitzscaling Ventures general partners score companies collectively, to engineer the kind of rigor that prevents this kind of manipulation.  Many times a month, we will start scoring a company that we love, perhaps because the product resonates with us, or because we love the team, but we find after scoring just the first two categories that we have to disqualify it.

  6. The one exception we make is that for companies that are close (e.g. 77-79) where at least one of the partners is a strong advocate, we are willing to start the process of deeper due diligence in case that uncovers new information that would lead us to revise our initial score into the 80+ range.


For startups that want Blitzscaling Ventures to invest, the formula is simple.  Convince us that you’re targeting a market with a strong winner-take-most dynamic, and articulate a scalable distribution strategy that will allow you to outgrow the competition and win that market.


For example, Airbnb would score well for the following reasons:

  • It is a two-sided marketplace; hosts want to be on the platform with the most guests, and guests want to be on the platform with the widest selection of properties.

  • It is a high-value; high-consideration transaction–no one books the first Airbnb they see in the search results; you’re spending hundreds of dollars and sleeping there, so you really want a good selection.

  • Hosts have a strong incentive to market their properties, both on and off the platform, to drive revenue and positive reviews that make it easier to earn future revenue.

  • Hosts can become guests and guests can become hosts, adding to the virality.


These factors give Airbnb strong scores in both winner-take-most and in scalable distribution.  From there, it has demonstrated great product-market fit, is tackling a huge market, earns a nearly 100% gross margin on marketplace revenue, and scales reasonably well (though its safety and guest experience operations, including arranging photographs of properties and responding to urgent customer service issues mean that its scores on organizational and operational scalability are far from perfect).


You can see these patterns in Reid Hoffman’s own career; PayPal is a classic example of a payment network with strong network effects, and incredible virality (both organic and incentivized).  Same for Facebook (especially with its college rollout strategy, where it leveraged college students’ existing friend networks to grow within colleges and spread between them).


And it’s no surprise that the world’s most valuable companies score well in the elements of blitzscaling as well.  Operating systems (Microsoft, Apple) and cloud computing platforms (Amazon, Microsoft) all have strong network effects, as does search advertising (Google) and social networks (Facebook).  This is true even in AI, where OpenAI has clearly won most of the initial market for foundational models, and Nvidia has done the same for GPUs.


Bear in mind, however, that the Blitzscaling Scorecard is powerful, but it isn’t magic.  Just because a startup is a potential blitzscaler doesn’t mean that it is bound to succeed and make its investors a ton of money.  Just because a startup isn’t a potential blitzscaler doesn’t mean that it can’t execute brilliantly and be a massive success as well (we often cite the example of Snowflake as a company that shouldn’t blitzscale, but could and did scale conventionally into a massive success).


But if you believe that the goal of a startup should be to win a valuable winner-take-most market, the Blitzscaling Scorecard can be a vital tool.


If you’re an investor like Blitzscaling Ventures, you can use it to identify high-potential startups that you can then engage with on a traditional due diligence process.


If you’re a founder, you can use the scorecard as a guide to improve your startup’s blitzscalability.  Maybe you need to build more network effects into your product, or maybe you can find a partnership that will supercharge your distribution.  And even if you’re still in the earliest stages of your startup, even if it’s just an idea on a napkin, figuring out your blitzscalability up front can help you focus on maximizing the upside of that idea so that you can improve your chances of building a massively-valuable business with lightning speed.



1,023 views0 comments

Recent Posts

See All

Comments


bottom of page