Are You Ready to Make the Decision?

Leading a startup means making important decisions with incomplete information. Achieving startup success demands relentless forward motion, so when do you have enough to make a bet?

Having certainty when making important decisions is often a luxury that a startup leader cannot afford. Achieving certainty typically requires some combination of time, data point repetition, and comprehensive contextual knowledge, all of which are in short supply for fast-moving startups. This means that a critical startup leadership skill is being able to find the sweet spot where you have enough information to make a good educated decision, but not waiting for certainty, which will usually require too much time and resources.

Here are a few examples of “big bet” decisions I had to make as a startup CEO, along with some of the factors, both known and unknown at the time, that factored into the call.  Hopefully, these stories can help illustrate some of the balance inherent when you need to make a decision, even while some factors remain unknown.

• Starting a beta test: Preparing for a product launch and eager to show good results to support our fundraising, we were looking to gather customer feedback on our life science tool beta prototypes. Our tests had shown promising results with cellular analysis, which was what our instrument had been designed for. A well-established assay company had volunteered to beta test our instrument with some of their bead-based assays. We were excited to get new insights from a different potential customer segment with a different type of assay that we had not tested before. In parallel, we had a Series B term sheet for our first venture-backed round and were working with our lead VC to recruit another VC to fill out our syndicate. One of the lead VC-candidates just happened to be a backer of our prospective beta tester.  It seemed like a double-win, and I decided to go ahead with the bead-based beta test.
  
  Two months later, that call proved to be a costly mistake. To our dismay, we discovered that beads were much more demanding on CV (coefficient of variation) measures than cells, resulting in unacceptable instrument performance on the beta tester’s assays. The prospective VC heard of our failure, and consequently pulled out of the emerging syndicate with a negative report to our lead investor.  As a result, I ended up in an ugly conversation in which the lead VC slashed our Series B valuation in half, putting us into a down-round situation. Ouch! In hindsight, I wished I had realized that the performance requirements for beads were higher than for cells, which we could have discovered had we thought to ask some of our scientific advisors. I wished that I had done some internal performance validation before jumping straight into a beta test situation, especially a high-stakes one like this. Of course, if it had gone well, we would have landed an impressive VC, an up-round, and a likely customer, which was the opportunity that enticed me into taking the chance.
   
• Launching the product:  With our investors’ support, we made the strategic decision to ensure that we had worked out the kinks in our product before a formal product launch through an extended beta test period. With a research tool, we had the opportunity to pursue agile product development, with fast iterations to ensure that we were satisfying our future customers with real-world limited-time beta tests. We did 25 beta tests in 8 months, with rolling improvements. Towards the end of the process, as we asked our beta testers to return their prototype devices, we started getting asked if they could buy the prototype. At that point, I made the call to launch the product commercially.
  
  With the benefit of hindsight, we did make the right call going to market when the customer pull became clear and apparent through the desire to buy a beta prototype rather than give it back to us. When I made that call, we certainly didn’t know everything. We had additional improvements in development (within a year, we had developed an easier-to-manufacture, lower COGS instrument) that I could have thought were necessary before launch.  Instead, when we gave our last beta-testers the opportunity to buy their prototypes at a discount, they started referring others to us. We were able to initiate a virtuous cycle of selling and manufacturing finished devices in response to an emerging list of purchase orders.  The critical decision factor was making sure that our beta tests were always for a limited time, which encouraged our testers to decide that they didn’t want to let their device go once it was consistently meeting their needs, which gave us the clear signal we needed to make the call.
    
• Shutting down a startup: Our world-leading team of physicists and engineers had developed an incredibly novel and advanced technology that was intended to radically reduce the cost of utility-scale offshore wind power generation. We had raised equity capital and secured $5M in support from the U.S. Department of Energy’s Advanced Research Projects program. All of our backers were excited about our progress – and yet, gradually, the world began to change. Using the status quo utility-scale wind turbine technology, the European wind industry started collaborating across organizations to drive 50% out of the cost of developing an offshore wind farm. These competitive landscape signals took years to emerge and confirm, but that trend ultimately wiped out our anticipated competitive advantage in an incredibly cost-sensitive industry. At the same time, our critical next milestone – deploying our technology in a real-world offshore prototype – kept being forced to shrink based on out-of-our-control constraints that kept evolving, always in the wrong direction. And there was emerging clarity on physical to our team members, in the context of an eroding potential benefit. In the end, despite continuing support and pressure from the Department of Energy, our existing investors, and our team, I made the call to close down the company and return capital to our backers because our potential value proposition was likely wiped out, even perfect performance on our next milestone (never realistic) would not support the next level of fundraising, the diminished upside could not justify the risk. It was a hard decision at the time – and was opposed by most everyone around me, yet years later, the murky factors that went into it were proved to be certainties – and the call was confirmed to be the correct one.   
       
• Submitting to the FDA with unproven performance targets: Engaging the FDA is always time-consuming, and time-consuming is costly both in terms of a startup’s burn rate and investor’s confidence.  As we sought to advance our conversations with the FDA on a truly novel technology, we needed to declare our acceptable performance targets for an unusual clinical study. We had some preliminary data, but it wasn’t collected under the same conditions as what we were proposing to FDA.  Ideally, we would have run some experiments to confirm our targets before declaring them, however, at the time, we had discovered that the data that we had available and were continuing to collect had a newly discovered input quality problem. This systemic data problem meant that we could not rely on our regular data sources to provide valid results until those problems were resolved, an uncertain process that had already been underway for months and was not likely to be resolved for quite a while. The question I faced was do we draw a line in the sand based on imperfect, older data or wait another six months or more to determine if we could more confidently define the goals?  Six months of delay would likely demand an onerous bridge from our investors. I held my breath – and submitted the best goals we could come up with as we tried to balance being as conservative as possible while still achieving a sufficiently good performance level to justify FDA positive action.
  
  Ultimately, we resolved the data quality issues, and had a more stressful pivotal clinical study than we had planned on when our targets turned out to be more demanding than we realized due to some statistical nuances that we did not yet understand. In the end, we had a successful study that exceeded its endpoints – and we did not have to raise money under duress, so the decision proved to be a good one.  It could, however, have gone the other way had we not been able to produce a successful study.

Making decisions in the gray space of the unknown, using the best available information, is often essential to advancing a startup successfully. Sometimes those bets do not work out as hoped. Sometimes they do. Strong planning and synthesis skills are often the foundation on which such decisions are made – and, over time and across many experiences, hopefully, you can tune your sense of risk so that you make the right call often enough to make up for the inevitable errors in judgment or just bad luck that happen along the way.