# 🎙️ Lenny's Panel: Executive / Board Review **Heat:** 🌶️ Heated (4/5) **Panelists:** Adam Fishman, Bobby Pinero, Elena Verna, Shreyas Doshi **Document:** "[DRAFT v1] Fund the Onboarding Rebuild — Approve a two-quarter rebuild of new-customer onboarding, staffed by 6 engineers, to cut time-to-first-value and stop the activation losses at step 3." > **Grounding note:** panel selected from `disagreements.json` (295 relevant pairs) — Fishman and Pinero hold documented positions that directly contradict this doc's core premise. Frameworks and views below are pulled from their archive profiles. --- **ELENA VERNA:** Let me start with what's missing, because it's structural. This doc has a target — "2-minute first value" — and no growth model behind it. In my **Growth Model Framework**, you pair growth levers with growth motions so someone is accountable for the outcome. This says "rebuild onboarding" and then asserts activation improves. That's a project plan, not a model. And the metric hierarchy is upside down. I think of **data as a company org chart**: revenue is the CEO, acquisition/retention/monetization are the leadership team, activation and conversion are team members. This doc has promoted a team member — setup time — to CEO. Your CFO is going to ask what revenue does, and "11 minutes to 2 minutes" is not an answer. **ADAM FISHMAN:** I'll go further, and this is the part that worries me most. The doc treats the drop-off at step 3 as pure loss. It isn't necessarily. **Good onboarding should sometimes decrease conversion if it improves retention by filtering out unqualified users.** Identity verification is a filter. Some of that 40% is people who were never going to be good customers — and if you defer the check to get them through the door, you don't get activation, you get a worse cohort further down the funnel plus a compliance mess. Nobody has segmented that 40%. Who are they? Do the ones who *do* push through step 3 retain better? Until you know that, "recover a large share of the step-3 drop-off" is a wish. **BOBBY PINERO:** Agreed, and I'd put it more bluntly because I've built the opposite of what this doc proposes. My **Strategic Friction Principle** — friction in onboarding can *increase* activation, because commitment mechanisms stop people from skipping the setup steps that actually make the product work. At Equals we require a data-source connection. It's work. It's the reason people activate. The doc's central move is "reach first value *before* the heavy identity check." That's the exact instinct I'd challenge. **Removing setup steps that seem difficult often hurts long-term user success.** You may be optimizing the number that's easy to see — signup completion — and damaging the one that pays you. **SHREYAS DOSHI:** I want to name the real decision, because the doc buries it. It presents this as "rebuild vs. incremental." That's not the question. The question is **opportunity cost**. My framing is: stop asking "is this a good use of resources" and start asking "**is this the best use of these resources**." The doc admits it slips the reporting API by a quarter — an API committed to *two enterprise accounts* — and then spends one clause on it under "What it costs." That's the whole ballgame for a CFO. You are trading a committed enterprise obligation for a modeled activation gain. The doc doesn't quantify either side. It should not survive contact with a finance team in this shape. **ELENA VERNA:** And the competitor number is doing unearned work. "They shipped a 2-minute setup" — so what? Different segment, possibly different regulatory exposure. If they're not doing KYC at signup, they aren't running your business. Benchmarking your setup time against a company with a different compliance surface isn't strategy, it's anxiety. **ADAM FISHMAN:** Which connects to something I feel strongly about: **don't redesign onboarding frequently — only when you have genuinely new insights about customers.** What's the genuinely new insight here? I read an offsite where everyone agreed it feels broken, one unverified funnel number, and a competitor's launch. That's not insight, that's stimulus. Where I *will* defend the doc: onboarding is the only product experience 100% of your users touch, and it's the first delivery on your brand promise. Under-investing there is common and costly. So the instinct is right. The evidence is not. **BOBBY PINERO:** There's a cheaper version of this that the doc never considers, and it's the one I'd fund tomorrow. Treat **onboarding as sales** — it's a process for convincing someone to continue, not a corridor to walk down. Most of that is sequencing, copy, expectation-setting, and support at the moment of friction. Very little of it requires six engineers for two quarters. If the identity check is the wall, the first question isn't "how do we move the wall," it's "why does hitting the wall make people leave?" Do users understand *why* you need it? Do they know how long it takes? That's a week of work, not a rebuild. **SHREYAS DOSHI:** Run a **pre-mortem** before you spend a dollar. It's two quarters from now, the rebuild shipped, activation didn't move. Why? I can already write three of the answers: the drop-off was qualification, not friction; progressive verification created a compliance escalation nobody modeled; the 40% was measurement error. **Pre-mortems beat post-mortems because they prevent the problem instead of teaching you about it afterward.** This doc has a "Risks" section listing two risks it has already dismissed — "compliance: we'll defer it," "migration: unaffected." Risks you've pre-solved in a sentence aren't risks, they're reassurance. **ELENA VERNA:** One more thing on the ask itself. Six engineers, two quarters, one shot. Why is the unit of investment a rebuild rather than a sequenced set of bets against the model? If the model says activation is the binding constraint, you fund the cheapest test of that first. If it isn't the constraint, you've saved two quarters and an enterprise commitment. **ADAM FISHMAN:** Here's my constructive version: instrument before you build. Segment the step-3 drop-off by cohort and look at retention of those who complete it. Then use **opinionated defaults** — make it hard to do the wrong thing, easy to do the right thing, without stripping the step out. That's a scalpel, and you'll learn whether the big investment is warranted. **BOBBY PINERO:** And if the data comes back saying the friction is qualifying rather than blocking — you will have avoided burning two quarters to make your funnel look better and your business worse. **SHREYAS DOSHI:** Which is the sentence the doc needs to be able to survive. Right now it can't. --- ## Board Memo **The strategic question.** Leadership is being asked to approve six engineers for two quarters to rebuild new-customer onboarding, on the thesis that friction at identity verification is destroying activation. The panel's central objection is that the doc asserts a causal story it has not evidenced: it treats a ~40% step-3 drop-off (self-described as unverified) as recoverable loss, when a material share may be qualification — users filtered out who would not have retained. Two panelists hold documented positions that friction can *improve* activation and retention, making this the crux rather than a footnote. **The options.** (1) **Fund as written** — highest cost, highest regret risk; the panel does not recommend it on current evidence. (2) **Instrument first** — segment the drop-off, measure retention of users who complete verification, and test comprehension/expectation-setting at the friction point. Low cost, weeks not quarters, and it either kills the thesis or arms it. (3) **Fund a scoped slice** — sequencing, copy, and defaults around the existing check, without deferring the compliance gate. The panel converged on (2), with (3) running in parallel. **Recommendation and risks.** Do not approve the two-quarter rebuild in its current form; approve an instrumentation-and-experiment phase with a decision gate at the end of it. The binding constraint on approval is not engineering capacity — it is **opportunity cost**: the doc trades a reporting API committed to two enterprise accounts for a modeled, unquantified activation gain, and quantifies neither. Milestones before re-proposal: (a) verify the 40% and segment it by retention, (b) state the activation gain in revenue terms, (c) price the enterprise commitment being delayed, (d) run a pre-mortem on the compliance-deferral design. Re-present with those four and this becomes a fundable ask. --- *The panel did not weigh in on the rewrite-vs-incremental engineering approach or the specific UI redesign — worth a separate session once the demand-side evidence exists.*