Solving Challenging Problems

Like most children who were growing up in the early 1980s (I’m dating myself fiercely), one of my life’s goals was to be able to solve the Rubik’s Cube. Sure, I could have easily bought a book that mapped out the various solution methods (a book was required since YouTube wasn’t around then!), but I wasn’t that type of kid. I had to do it on my own. Nothing else would make me happy.

And I mistakenly thought it was going to be incredibly easy because of the progress I made in the first few days that I owned my first cube. I figured out that the center square on each side never moved (obvious but important). I figured out how to complete one side with all the adjacent edges the same color as their center squares (quite easy). And within a week I figured out how to fill in the second layer on all sides so that only 8 pieces needed re-positioning to complete the cube (also quite easy). And then guess what happened?  I stalled out big time making zero progress for months. All my attempts to figure out what I thought were “just a few moves to finish the damn thing” were for naught. The cube soon became a paperweight and went unsolved (and untouched) for months.

The reality was that I didn’t know how to figure out the last series of movements because I didn’t have a method or a framework for advancing my understanding of the cube. The early steps were super easy to figure out and didn’t actually require much more than a basic understanding of how the cube worked. What I needed was a much deeper understanding about how to move the pieces and for that I needed to actually put in some real work. I eventually picked the cube back up (I wasn’t going to let it beat me) but this time with determination and focus. I did end up solving the cube (whew!) but only after I treated it as a mathematical puzzle vs. a simple toy.  My big breakthrough came once I set goals for what I needed to accomplish and mapped out solutions (i.e. – rotate a corner piece in position or move an inside piece to a different side). It required solving a series of smaller problems on the way to solving the bigger problem. And it required not stopping until everything was in its perfect position. There was real work involved but in the end it was worth it.  A solved cube was a thing of beauty and a true life accomplishment to my 10 year old self.

This same concept applies to the startup world in so many ways. Unfortunately, I’ve met too many Founders who stopped after solving the “easy 80%” or never bothered to think hard about the fundamental truths of the business model they were following as they built out their business. Too many business models that I’ve reviewed feel like they’re Rubik’s Cubes with one side and two rows completed. They feel “nearly complete” but in reality they’re not. They represent an unsolved problem.

And in the startup world, incomplete solutions usually have intrinsic risk associated with them that should be internalized.  Holes in a business model might be solvable over time….but they might not.  And it’s entirely possible that the one thing that hasn’t been solved is a critical driver of the business’s ultimate success or failure.  To be clear, Venture Capital is about taking on the risk of unknowns, but there’s no reason to build a business using an incomplete framework.  The risk of chasing the “easy 80%” is that a business might look like it’s getting fantastic traction but it could lead to a dead-end that’s unsustainable on any of a number of dimensions (i.e. – financial viability, regulatory concerns, availability of funding, consumer value and retention, stability through cycles, etc).

For instance, I’m personally quite intrigued with the very difficult problem of how to underwrite and structure loan products for workers in the emerging 1099 economy. A quick review of labor trends reveals that it’s becoming more and more common for people to generate at least some income from jobs that have optionality and variability associated with them. And for some workers, 100% of their income is generated “at will” vs. contractually. An Uber driver can work when they want to. A graphic designer might generate enough business to work 80 hours one month and 20 hours the next. If a 1099 worker wants to take a week off to go on vacation, they earn precisely zero while sunning on the beach.  And the sad but lurking truth is that many gig economy workers actually take work in a 1099 capacity as an expression of their economic fragility rather than an expression of their free-spiritedness.

So how do you underwrite and structure loans for these types of workers? A few of the businesses I’ve seen in the space only attempt to solve a piece of the puzzle or take underwriting shortcuts as approximations for what’s really going on. They’re starting with what’s “easy” and not digging deep enough to get to the fundamental drivers of how to generate the best outcomes for this segment.

Today’s reality is that loan products (and systems) are designed around the concept of stable income backing a loan schedule with routine and fixed payments (with credit cards having a bit more flexibility given the tiny minimum payment requirements). Traditional underwriting models attempt to pull stability of income and willingness to pay signal out of historical payment patterns (I.e. – bureau data), calculate an applicant’s present ability to pay (i.e. – ratios of income to debt and free cash flow calculations) and might go so far as to evaluate savings and/or collateral as fall-back in case there’s a shock to the system (I.e. – major medical bill or temporary job loss).

But these traditional methods fall apart when work is variable, optional and in many cases unstable. New, very unfamiliar concepts need to enter into the system if underwriting is to be done properly. If one could determine “willingness to work”, “ability to generate work” and “ability to plan for periods with little/no income” it most likely would explain quite a bit of the variance in loan performance within today’s 1099 worker population. And if the product itself could be structured to allow for periods of reduced/no income generation then it should increase the repayment rates for lending products. One of many challenges will be designing a product and underwriting methodology that works for debt lenders, consumers, regulators and charter Banks, but this is an example of what designing a complete solution is all about.  Imagine a mortgage product that allowed a borrower to take a week off (vacation) and only pay 75% of their scheduled payment that month without being deemed delinquent. Or a product that would allow a borrower to over-pay in good months to reduce their payments in months where work is a bit harder to come by.

I just haven’t (yet) discovered a business model in the space that solves 100% of the Rubik’s Cube. And the businesses I’ve seen operating in the space (including incumbents) seem very satisfied with their 80% solution. You’d be shocked at how some of the biggest traditional mortgage and auto lending originators are approaching the problem (mostly ignoring it and assuming their current models work) and you’d also be equally amazed by how even the most innovative startups are relying on traditional underwriting theory and product structures to tackle what requires a new way of thinking. Many times their products and methodologies are being justified by the crutch (and truth) that their debt providers or bank charter providers or internal credit teams don’t want to stray from what they “know”.  Wouldn’t it be great to know that you’re heading down a path that could lead to a complete Cube vs. one that almost by definition will leave a few pieces in the wrong place?

That’s what we look for in Investments at QED. It’s why we have yet to make an investment in spaces like “shared equity mortgages” and “decumulation of assets” even though there are quite a few talented teams tackling these super interesting problems. We’ve seen plenty of “good” business models….but we’re looking for frameworks that could lead to complete Cubes and we enjoy helping Founders find solutions when the last 20% isn’t obvious.