Startup Mortality: What End-of-Life Care Teaches Us About Startup Failure

I’m shutting down Esper, an enterprise machine learning (ML) startup that I founded 4 years ago. We raised roughly $4M of venture capital and were used by clients at Dropbox, Salesforce, Stanford, Kohl’s, and BBVA Compass. At this point, it’s customary to provide a platitude like “it’s been an incredible journey,” but that would belie the difficulty of the decision.

A founder’s greatest strength is their resolve — their ability to “rage against the dying of the light.” Yet at the same time, their greatest weakness is that they sometimes just don’t know when to quit.

Sure, a good founder knows to divert resources away from something that isn’t working, but what happens when that something is the startup itself?

I recently read Atul Gawande’s Being Mortal, in which he discusses medicine’s role in end-of-life care. I found some striking parallels between end-of-life startup decisions and end-of-life medical decisions.

Startups start with an end-of-life diagnosis. As in medicine, we can make decisions to “prolong” the life of a startup. We pivot, receive bridge funding, or dive into a completely new market. When things don’t get better, we pray for a miracle (e.g., an acquisition).

When the end does come, you rarely hear the stories illuminating what could have been done differently. Despite the adage in Silicon Valley that “failure is celebrated,” you’ll be hard-pressed to find obituaries of Sequoia’s defunct portfolio companies. Instead, you find Bessemer’s anti-portfolio of “missed opportunities.”

In an effort to break the cycle, here is Esper’s story of failure. While I lay out our choices to prolong life and our attempt at a miracle, perhaps what is more important is the discussion around what we would have done differently.

Asking the right questions earlier, seriously reassessing our market, and finally, letting go. For those in a similar position, I hope this story helps you. The end does not have to be bitter, abrupt, and painful; there is a way to “go gentle into that good night.”

Esper: Our Choices

“Life is choices, and they are relentless. No sooner have you made one choice than another is upon you.” — Atul Gawande, Being Mortal: Medicine and What Matters in the End
Esper’s technology created top-tier scheduling by sending agendas, event invitations, reminders, and keeping notes related to tasks

For the first 2.5 years of Esper, we had about 20 employees/contractors and $2.7 million in funding. Esper started as a scheduling service (i.e., cc Esper’s outsourced assistant service powered by our technology, which would eventually automate all scheduling).

Unfortunately, we found that over 90 percent of the benefit in the service was created by people and not technology, a problem seen in many similar startups. As we scaled, we found ourselves frantically managing people rather than creating software, which cut into our already small profit margin.

And as we charged more in enterprise, customers left the service because they could hire in-house assistants instead. We just didn’t provide a 10X ROI over hiring an assistant. So we chose to pivot. We ditched the assistant service and focused solely on providing software to those assistants instead.

New Esper — Mint for Time

Esper downsized the following year to 8 employees. We gained a new CTO, received an additional $1 million, and moved into a converted Subway restaurant. Esper started delivering software for assistants, focused on calendar analytics for executive assistants and executives (e.g., Mint for Time).

Despite gaining traction with executives and assistants, we failed to achieve necessary growth amongst businesses.

First, we achieved limited viral growth with links in calendar invites and emails to the calendar invite guest list (in the form of meeting agendas and post-meeting ratings).

Second, when we focused on selling, we found difficulty selling our horizontal enterprise solution. Beyond the sensitive security hurdles (e.g., access to all calendar data), companies did not have a clear buyer. While finance software has a CFO, no company has a formal “Chief Time Officer.” Most buyers wanted to check a chart on a monthly basis, which didn’t demonstrate strong engagement in the product.

Finally and most importantly, our team started losing momentum with this new pivot. Whereas we were once excited for the grand vision of AI/ML scheduling, we were now just an incremental improvement on top of the calendar.

Though we thought were ready to take on this new “treatment,” the disappointments began taking a toll on the team.

We created tailor-made slide decks regarding team and technology for each acquirer.

We were now hoping for a miracle — an acquisition. As we went through this arduous process, we found interest amongst large AI/ML/Assistant companies, enterprise companies, and companies interested in our ML expertise. However, we soon learned that they wanted our team, not our technology.

And then the final blow — we lost the term sheet to our most exciting acquisition. It was at this moment that we asked ourselves: is it worth it to keep going? The questions we asked ourselves were very similar, I later learned, to those posed by Dr. Susan Block in Being Mortal:

What is your understanding of the situation and its potential outcomes?
What are you fears and what are your hopes?
What trade-offs you are willing to make and not willing to make?
What is the course of action that best serves this understanding?

Unfortunately, we asked these questions too late.

Go Gentle into that Good Night

“You may not control life’s circumstances, but getting to be the author of your life means getting to control what you do with them.” ― Atul Gawande, Being Mortal: Medicine and What Matters in the End

Being Mortal helped me redefine the concept of “failure” for startups. Failure is not choosing to shut down a startup; it is not having the agency to determine how the story ends.

As Atul Gawande writes, “our ultimate goal, after all, is not a good death but a good life to the very end.” When you’re contemplating the end, it’s important to step back and ask the right questions in order to create the most positive experience possible. Here are a few things I learned:

Ask the team what they really want before acquisition talks. I had discussed what folks were interested in at a high level, but I never asked them point blank about their priorities and tradeoffs.

I thought we wanted an acquisition, but had I held the right conversations, I would have realized otherwise. So I led us into acquisition talks. The process split our team apart for interviews — or what felt like interrogations — and took us on an emotional roller-coaster of uncertainty that quickly left a sour taste in our mouths.

Once you start the acquisition process, the timer begins, so it’s better to have a clear understanding of priorities and tradeoffs beforehand. I wish I had asked Dr. Block’s four questions to myself and to each team member earlier. It may have meant an earlier decision to shut down and a better life to the very end.

Regularly assess the potential of your market. Elad Gil recently said: “Go work at a company with a great team and terrible market for 6 months. Realize that markets sadly matter more than teams in terms of success…”

Working on a startup in a bad market may require more energy than one in a good market. Looking back, like a patient learning their prognosis, I wish I understood the contours of our market sooner.

In the world of calendaring and assistants, there may still be an opportunity to build a vertical software startup for high-end assistants (e.g., personal/family assistants of high net-worth individuals).

Otherwise, it’s unlikely that the solution is a 10X ROI beyond a feature in the calendar or a messaging channel — there’s just not a big wave occurring in productivity software. Had I realized this sooner, I may not have pursued a bridge round.

Don’t hesitate to reach out. When startups were scarce, VCs used to intervene (sometimes against the founder’s wishes). Nowadays, a plethora of startups have left many founders feeling alone.

While this has given founders more freedom, it’s also meant less support. Unlike end-of-life care, there is no hospice or assisted living for startups (e.g., a place where dying startups can receive support on job placement for team members, acquisitions, or legal matters).

Nevertheless, everyone (not just investors) is willing to help. Although I felt embarrassed and guilty, once I reached out, I realized how common my situation was and how empathetic others were.

I wish I had reached out sooner. This startup ecosystem is critical to resource stewardship, ensuring promising startups have the funding they need, and maintaining a healthy startup culture so that founders and employees aren’t burnt out on the industry as a whole.

Learn to let go. For some founders, the decision to move on is easier (e.g., they draw a clear line in the sand around economics, team, or timeframe). For me, the decision was harder.

I’ve always admired the captain who goes down with the ship — the Old Man and the Sea, Battle of Thermopylae, and every other “guns ablazing” scenario. Whether driven by indefatigable pride or fear, I was willing to hold on relentlessly, clinging to the mast until the bitter end.

In fact, I thought that was what founders do. What I had was courage, but what I needed was wisdom. In the fog of war, you’ll never have perfect information, but that doesn’t mean holding on relentlessly is the best path forward; sometimes, disbanding may be the most prudent option.

In dealing with death, Dr. Gawande aptly uses the following war analogy:

“…you don’t want a general who fights to the point of total annihilation. You don’t want Custer. You want Robert E. Lee, someone who knows how to fight for territory that can be won and how to surrender it when it can’t, someone who understands that the damage is greatest if all you do is battle to the bitter end.”

Consistently, every startup veteran (investor and founder) echoes the sentiment that failure is not the worst thing; the worst thing is working on something for years with no end in sight. While I’ve always known this statement (in fact, I’ve repeated the learning for many others), only after having gone through Esper did I truly understand the value in a startup’s mortality.