This post originally appeared on Built in Chicago.
Much of my job consists of reading about companies our fund might want to buy. I usually read about them long before I meet the people who started them. I have a soft spot for entrepreneurs, as you might already know, so I am easily inspired by all the different ideas people have turned into businesses. However, when you do something a few hundred times a year for a couple decades, you can get pretty surgical about what you look for in a new investment. Often this means the stories behind the businesses (and the people who built them) get lost in the routine because when you are working through a stack of investment opportunities you don’t make time for nuance. I’m sure our limited partners feel the same way; they spend all day reading about fund prospects and their “value-added, proprietary strategies” and their “top-quartile” returns. Even trying to avoid cynicism, it’s easy to get a little jaded over time as you look for the uniqueness. As a result, we develop heuristics to speed up the process.
For anyone who has read Daniel Kahneman’s book Thinking Fast and Slow (here’s a review so you don’t actually have to slog through it), you know these heuristics are just our brain’s way of avoiding the hard work of thinking through complexity, an automatic response our brains use to find shortcuts to the easy, comfortable answers. A recent paper by a team at Georgia State University utilized research about heuristics by Kahneman and others and applied it to how private equity (PE) firms screen for new investments. For the most part the paper is pretty “inside baseball” stuff, but if you’ve ever worked in the industry the quotes from the PE partners they interviewed for the project could easily have come directly from the Monday deal meetings at any fund in the country.
Even if you’re not into behavioral economics books or academic papers on PE, if you spend any time around people in the PE or venture capital industry, you will hear the term “pattern recognition” all the time. “This is a pattern recognition business” is definitely in the top ten PE clichés, just behind “value add” and “proprietary deal flow.” This term is generally shorthand for “I’ve been in the business for a while; I have a lot of experience; I’ve seen it all; nothing gets past me.”
Some of the common “patterns” in PE are: no husband and wife teams; no first-time CEOs; never get involved in a land war in Asia. These simple statements are written in stone at most firms, as the Georgia State authors found out, and pretty much every partner at every firm has a few extra of their own handy for every investment their fund considers. Most PE execs would posit that having pattern recognition gives you a leg-up on the rookies who haven’t put in the required time to acquire this “mystical power” themselves. “You aren’t given this power from a unicorn, son, you only acquire the right to be intellectually lazy after years of hard work! Now print me out that board deck so I can read it on the early train to Greenwich!”
In this industry, your pattern recognition skills are a badge of honor. Much of your conviction for, or against, a deal is built on the patterns you’ve seen fail or succeed in the past. But what is meant to be code for rigor, experience and wisdom is actually just Daniel Kahneman’s pesky “System 1” thinking fooling you with false confidence. In reality, the patterns private equity folks see around them aren’t patterns at all; they’re just mirages of simple thinking which your lazy brain defaults into so it can avoid taxing itself with complexity.
Here’s my simple case against these pattern seers: say you’ve been in the business for twenty years (which is pretty long for a relatively young industry like private equity) and say you have closed one new investment each year during that time (and that would be a hall of fame pace – I’m not joking there is actually a PE Hall of Fame). This would give you twenty data points across twenty different companies. But these companies operated during different economic cycles, were led by different management teams, and competed in markets with different macro and micro forces driving them over the years. Of course each company also experienced varied growth, regulatory, competitive and other critical dynamics during the investment period. Because of all this variability, I don’t believe that you can really develop any reliable patterns here.
Even if you include all the investments made by everyone at your firm during that same period, included a few others from friends’ firms, a handful of stories you heard where investments didn’t go so well for one reason or another; I’d still argue there isn’t enough data to discern anything close to a pattern – there are just too many competing variables to solve for in such a small data set. At best, you might have some metadata; but that’s a topic for another post.
Instead, what you have is Kahneman’s and Amos Tversky’s (a long time collaborator of Kahneman) Availability Heuristic. This is a mental shortcut people use to predict the probability of something because of how easy it is to think of examples combined with a heavy dose of Substitution (using a simple question to answer a difficult one). Everyone who has ever sat through a Monday deal meeting at a PE firm knows what I’m talking about. Invariably, someone will bring up a recent personal experience to make a point about a potential investment and expect that to count as reliable evidence. In the Georgia State paper, the authors found similar behavior across a couple dozen firms using Paul Slovic’s (and others’) Affect Heuristic – the reliance on feelings when making decisions especially when judging risks.
Calling it pattern recognition makes it sound like there is data and math behind the argument (the pattern part), wrapped up in the wisdom of a long career (the recognition part). It’s one of those things that is taken prima facie, without rebuttal, and it’s a skill younger practitioners in our business aspire to gain over their careers as a common knock on someone’s investment judgment is that they haven’t yet developed enough pattern recognition. But the idea of pattern recognition in private equity investing is classic “System 1” thinking; it’s dominated by emotions, not logic.
According to Kahneman’s research, “System 1” thinking uses the automatic, subconscious part of the brain and it’s where this thing called pattern recognition in PE thrives. Your brain spends most of it’s time using “System 1” to keep things simple, but there is plenty of scientific research proving how poor “System 1” is when assessing risk and seeing reliable patterns. The insidious part about “System 1” thinking is this: because it is automatic, it feels like it comes from wisdom and experience (your gut!) when it actually comes from laziness. Folks who study how the brain works recognize this for what it is.
By creating these heuristics our brains don’t have to fire up “System 2” to sort things out the hard way; with slow, effortful and deliberate consideration. This means the pattern recognition you summon isn’t the result of thinking deep thoughts and experiencing the hard knocks of a long investing career; it is really just your brain sitting in front of the TV with a beer in your hand, Cheetos dust on your stomach, watching Night Court reruns.
In an effort to practice what I preach, I’ll put my heuristics on the table in the list below. Like a much more eloquent Irishman than me once sung I’ll where it proudly. With my “System 1” thinking in full force, these are the biases I run through as I think about new investment opportunities:
Maniacal focus on the customer: Is the management team obsessed with the customer experience? Have they built a product that is the result of deep customer insight? By this I don’t mean that they “know their customer,” I mean do they do the things the great product companies do? See Scott Cook and the constant re-imagining that goes on at Intuit for a reference point. Have they kept this discipline as the company grew or does it need a refresh?
Subscription or transaction-based revenue model: Key points here are customer and revenue retention from year-to-year. For subscription models, a simple test to track momentum is to look at growth in the Deferred Revenue line on the balance sheet. And in case you were wondering, repeat customers don’t count as recurring revenue!
Mission-critical product: Is this something that the end user can’t operate her business without? Is it painful to switch? What are the available substitutes? That doesn’t necessarily mean that the product is a big piece of enterprise software. Just look at how payment gateways have become a critical component of the SaaS and mobile app market – pretty simple software but you can’t get paid without it.
Short sales cycles: I prefer companies who can close a new customer in less than forty five days (from first interaction to collecting the cash) and in some cases close a customer on the first interaction. Call this a “one call close.” Short sales cycles also generate lots of data so you can quickly adjust your approach and get immediate feedback. My goal with every company in which I invest is to build the talent and the tools to know if we had a good day or bad day every day.
Decision-maker contact: I like companies who can talk to the decision-maker in the sales process (reinforces my one call close above). I stay away from companies who sell to buyers with layers of decision-making or who sell products that require layers of approval. This drives my interest in companies that sell into the small-to-mid-sized businesses (SMB) and companies with discreet products that don’t require long evaluations.
Lots of targets: Maybe one of the more obvious ones on the list, but the market should have lots of end-customers to target. Even in small markets, you can build successful companies if there are a lot of customer targets. The key in a small market is you need to build the company that can win 50 percent or greater market share.
Cross-sell/up-sell: It’s always great if you can sell more products to your existing customers. Not all companies have this opportunity, but when they do, it can really drive predictable growth when combined with highly tuned organic sales efforts. Some call this “land and expand” and it can really work well in technology companies if you have a wedge application that gets you in the door and then build off of the great customer experience you deliver.
Cash flow: Subscription and transaction revenue models often lend to strong cash flow because your customers pay up-front, or because the revenue is highly predictable. At the stage our firm invests, we are well beyond the cash-burn days of startups; the companies we pursue often have very attractive cash flow margins. We’ll save the debate about what’s more important cash flow or earnings before interest, taxes, depreciation and amortization (EBITDA) for another time (but let me give you a hint, it’s cash flow!).
Non-linear upside: Are there elements about the market dynamic or the business model that could drive exponential growth? Like when you not only grow by adding new customers but you also grow when your customer grow (see the payments industry). Are there scenarios where a strategic acquirer would pay an outlier multiple of revenue or profits to own the business? Having these characteristics isn’t always necessary for success but they sure are nice to have.
Downside protection: This is the opposite of the point above but even more critical in my view. I like investment opportunities where, even if things don’t go as planned, I can still achieve a return through a good structure, an attractive valuation or a resilient business model. Private equity (unlike venture capital) is a business where your bad investments still need to return capital if you expect to generate decent returns for your investors. Losses are fund killers. As an investor in the rapidly changing technology sector, this means that risk mitigation is my prime consideration upfront.
As you can see from the list above, a lot of my biases revolve around product, revenue model and sales cycle. When it comes to analyzing these components for a new investment, call me old school, but I rely on primary research instead of pattern recognition to drive my decisions. Pattern recognition is the mental algorithm to avoid the hard work of fighting your biases to uncover great investment opportunities – you build it, it gets smart and then you just let it run. The difference in approach is like the difference between the algo/flash traders on Wall Street and the buy and hold guys from a generation ago.
While no company checks the box on every point above I look for varying degrees of emphasis in each of these components to quickly gauge my interest in digging deeper. Laying out my biases with my team helps expose where we might be missing something important or worrying too much about something unimportant to the likely outcome. I ask my team to throw theirs on the table along with mine to make the exercise transparent. This is where the deliberate consideration comes in and when the real analysis begins, absent any claims of pattern recognition.