Hey Reader,
In the last 7 days, what's the worst experience any user had in your product?
Not a Sentry alert. Not a bug you patched.
I mean the worst real thing a user actually experienced.
Do you know? (if not, then observability)
And if you're pre-launch.. What do you think it will be?
Was that a one-off... Or does that happen to 1 in 20 people?
Your product isn't judged by the best experience.
It's judged by the worst experience that happens often enough to matter.
|
|
Reddit Ads Outperform Other Platforms.
Tap into Reddit’s 450M+ highly engaged communities where your ideal audience actually hangs out. Beyond the noise of LinkedIn or Google.
Scaling ads or testing audiences? Reddit’s self-serve platform gives precision targeting and cheaper CPMs than traditional channels.
|
|
You already think in P95 terms, you just don’t call it that.
When you say 95% of users complete onboarding are you ignoring the 5% who don’t. That 5%? Are they really edge cases?
P95 is your disaster line:
- The point where 5% of users fall below acceptable.
- Not theoretical. Not rare. Not worth ignoring.
If your P95 API response time is 5s:
- 95% get a response in 5s or less.
- 5% wait longer, maybe much longer.
- They think your product is broken and they quietly leave.
P95 thinking is baked into how founders talk about their products in every dashboard. Every investor update. Every #BuildInPublic post on LI.
P95 is your disaster line. Actually. You choose your disaster line. What level of user pain are we willing to tolerate and still call the product good enough?
Thats a company decision, a strategic one, it's the minimum experience you're okay being judged by. Your line. Your standard. The floor your startup stands on, whether you choose it or not.
If it’s P95 then thats the threshold where 95% of users stay above (good experience), and 5% fall below (disaster).
P95 is where disaster begins.
The point I am asking you to consider, is with all our founder delusion. Ship fast, live in the product, test the happy path, celebrate the users experience, it’s so easy to ignore users falling below the disaster line without even knowing it. 95% succeeding seems amazing but 1-in-20 failing, somehow hits different.
And this compounds. Your learning comes from the clean cases of happy users having a good experience.
Your P95 isn’t a technical problem. It’s a business problem disguised as a technical problem and it ruins CAC. You spend $200 acquiring a user. If 1-in-20 hit your P95 experience and churn, you blame the usual suspects.. ICP mismatch, onboarding, timing. But often, your product failed them at the worst moment, when it mattered (to you and them!). And it’s often your best potential users falling below the disaster line. The ones who push limits, have complex setups, power users…
I had it today on an ecomm site. Their checkout pushed me into a flow on Safari with autofill enabled. I couldn’t fix it. Couldn’t check out. So I quietly bounced. They’ll never know they lost my order for a fixable reason. I’ll just show up as another “abandoned cart.” But I didn’t abandon. I hit their disaster line.
For my rage emailers.. P95 disasters are worth fixing. P99 disasters aren’t.
P95 isn’t about edge-case ghosts like some guy on a VPN in rural Kazakhstan running Linux and an adblocker and can’t download the PDF. That’s P99. Where we choose not to fix because the value isn’t worth the cost.
P95 is real users at real scale falling into disaster territory. The goal is to move your disaster line. (even just slightly, to a better place.)
Move the P95 experience. Now only true edge cases have disaster. So 95% of users succeed on their first try.
So, if you kinda like this concept, you should find your disaster lines. Track your P95 data. Page loads. Onboarding completion times? API response times?
See how many people are falling below each line and see if you can move this disaster line from catastrophic to merely annoying. From disasters into manageable problems?
That 1-in-20 disaster. It’s probably happening to someone right now. While you’re reading this.
If I can be of service, feel free to grab time.
LFG.
-- James
Not A Subscriber?
Join 140k+ StartUp Founders reading my weekly newsletter offering tactical insights to start, scale, and fund their startup. Real advice from a 3x exited founder.
|
8424 Santa Monica Blvd, Suite A523, Los Angeles, CA 90069
Unsubscribe · Preferences