The Sales Motion as Source Code
Treat the pitch like a codebase: run cohorts of 10, patch the bug the prospect found, redeploy.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 92%
Rather than running 100 sales conversations and then asking whether it worked, Kazanjy runs prospects in cohorts of roughly 10 first meetings, observes exactly where each conversation stalls, and commits a fix to the 'source code' of the sales motion — a new objection-handling slide, a reworded discovery question, a changed demo beat — before the next cohort. The sales motion is a living artifact under continuous revision, not a script you write once.
Origin
Kazanjy's own framing, deliberately borrowing the product-management cohort-release mental model (he pitches it to Lenny's PM audience as analogous to shipping a feature to cohorts of users and reading the funnel between releases).
Core principles
- 01A sales motion is everything you do to take a prospect through the process — and like software, it should be constantly updated.
- 02Batch-and-learn beats big-bang: 100 at-bats in one go destroys attribution.
- 03Every unanswered objection is a missing asset, not a bad prospect.
- 04Visual assets double as guardrails and a talk track for whoever runs the motion next.
How to run it
- 1
Prospect a cohort of about ten
Line up roughly 10 ICP-matched prospects and book 10 first meetings. This is one release cycle, not your whole pipeline.
Pro tip Keep the cohort tight on ICP so a failure is a message failure, not a targeting failure.
Watch out Doing 100 at-bats all at once and then asking 'did it work?' gives you a verdict with no diagnosis.
- 2
Instrument each conversation
Watch specifically: how the discovery questions land, whether they evoke the response you wanted, how the slides land, how the demo lands, and whether the prospect volunteers a next step (offering to introduce you to their boss is the good signal).
Pro tip 'I'd love to introduce you to my boss' is the cheapest, earliest proof your message landed.
- 3
Log the bugs
Note every question you did not have a good answer for, every objection with no artifact behind it, every place the narrative stalled.
Pro tip Write the bug down mid-call in the CRM — you will not remember it after four more meetings that day.
- 4
Commit a fix to the source code
Build the missing asset. If someone raised an objection you fumbled, make a slide that handles that objection so you can show it visually next time. Update the deck, the script, the templates.
Pro tip A slide is better than a memorised rebuttal: it gives you guardrails and a talk track, and it survives being handed to a rep.
Watch out Fixes that live only in your head cannot be packaged later — write them into the artifacts.
- 5
Redeploy on the next cohort and repeat
Run the next 10 with the updated motion and watch whether the previously-fatal objection now converts. Repeat dozens and dozens of times until the motion is repeatable.
Watch out Expect roughly two-thirds to two-fifths of conversations to still not work out even with a good motion — that is the shape of a healthy funnel, not a failing one.
In the wild
A prospect says the product doesn't sound like it does XYZ and the founder has no answer and no asset. The founder builds a slide addressing exactly that. Next cohort, when the objection recurs, the response is: a lot of people say that, but look here — and the prospect concedes the point and agrees to bring in their boss.
→ The sales motion and the collateral are both updated; conversion to second meeting rises for the whole subsequent cohort.
Common mistakes
Big-bang testing the pitch
Running all your prospects through an unchanged pitch and evaluating at the end wastes your scarcest asset — ICP-matched first meetings — and tells you nothing about which part of the motion failed.
Improvising the same fix every time
If you keep answering the same objection ad-lib without building an artifact, the fix never becomes transferable and the next rep will have to re-derive it from scratch.
Is it for you?
Best for
Founders and first sellers actively iterating an unproven pitch, especially product people comfortable with release-and-measure thinking.
Not ideal for
Mature sales orgs with a proven, high-volume motion where the constraint is execution consistency rather than discovery.
From the transcript
“you want to bring on like cohorts of users and then kind of like see what's going on then learn and then bring on new…”
“one helpful way to think about it is kind of like software, right? And like what you want to be doing is like constantly updating…”
“I'm going to update the source code, right? And like now my sales motion has been updated.”
“see how your conversations land see how the discovery questions are are evoking the right response or not the right response see how your slides…”
From the episode
Founder-led sales
Pete Kazanjy (Founding Sales, Atrium)