At EverQuote, we were growing fast. Auto insurance was working, the flywheel was spinning, and the conversation shifted to what's next. Expanding into home, life, and health insurance. A meaningful multi-vertical bet that would take 18 months to show results in any serious way.
The quarterly earnings calendar does not care about 18-month bets. It cares about this quarter.
I have done this at public companies and at companies heading toward public. The innovation challenge is not that people stop caring or that bureaucracy wins automatically. It is that the incentive system punishes the wrong things at exactly the wrong moments.
The quarterly clock is not your enemy. How you relate to it is.
The multi-vertical expansion that eventually took EverQuote from $200M to $400M in revenue did not start as a quarterly bet. It started as a protected R&D line item. A small team working outside the core auto business with a different success metric: learn fast, validate the model, then bring it to scale. That framing mattered enormously. The work was not competing with Q2 auto insurance numbers. It was running in parallel.
What I see fail consistently: treating innovation like a feature request. Loading it into the same sprint, the same roadmap, the same success metrics as the core product. At Spark Networks, we were managing a $50M media budget, quarterly subscriber targets, and three simultaneous platform migrations with the same team. Organic traffic was at 12% of industry benchmarks when I arrived. That is not a product problem. That is what happens when every team cycle is consumed by the quarterly number with nothing protected for fundamentals.
Compliance and process will take more than you expect. Build that in, not around it.
The gap between a good idea and a tested idea in front of real users gets longer after an IPO. Privacy reviews, security reviews, legal sign-off, data governance. These are not bureaucratic malice. They are the cost of accountability at scale. But they are also predictable.
The leaders who navigate this well map the process before they have an idea worth protecting. They know which reviews take four weeks and which take two days. They build shadow validation: small, low-stakes tests that generate real signal before the formal machinery engages. So when they walk into a steering committee, they are not asking for permission to test. They are presenting results.
At TripAdvisor, building the hotel e-commerce platform from zero to $200M in revenue in 18 months required exactly this. We were not fighting the process. We were running faster than it wherever we could, banking proof points, and using them to move approvals. You cannot prototype your way around legal review. But you can arrive at legal review with something real.
The innovation killer nobody names: the ad-revenue trap
I am working with a gaming company right now. Strong catalog, loyal 55-plus demographic, 20 years of institutional knowledge about their users. And a real problem: they know exactly what a great user experience looks like. Less friction, fewer interruptive ads, better retention mechanics. But they cannot get organizational alignment to take the short-term revenue hit to build it.
A new ad unit went live last year. It increased revenue. It also hurt the user experience in ways you could measure. Then YouTube launched a competitor product offering the same games for free. The competitive threat is visible. The response is not.
The unlock, when it happens, always comes from the same place: a leader who can model the recovery, not just the risk. Showing the board not just 'revenue will dip' but here is our 90-day path back to flat, here is what the retention curve looks like at month six, here is a comparable from a peer who made this move. You are not asking for faith. You are asking for a managed experiment.
What actually separates companies that ship from companies that announce
The innovation initiatives I have watched succeed share three things.
First, explicit portfolio separation. A dedicated line item, not a percentage of team time, but actual ring-fenced budget and headcount, for work that will not show results this quarter. This forces the real conversation about what the company is willing to bet on rather than letting it live in the aspirational language of an all-hands.
Second, different success metrics for different work horizons. Optimizing the existing product gets measured on revenue and retention. The next-bet work gets measured on learning velocity: how many hypotheses have we tested, how quickly, what did we find out. Conflating these two measurement systems is how you kill new ideas with old scorecards.
Third, leaders who can hold the line. Not because they are politically powerful, though that helps, but because they have done the work to model the downside clearly enough that the board understands what they are approving. Vague innovation requests die in approval processes. Specific, bounded experiments with modeled risk profiles move through.
If you are navigating this inside a public or pre-IPO company and want to think through how to structure the portfolio conversation, let's talk.
