Your session replay tools record everything. Yet the leak that's costing you money almost always slips past them. For three reasons: it's buried in the mass, it can't be reproduced, and every day spent looking for it has a price.
The reflex
Monday morning, the conversion rate is down 0.4 points. Nobody knows why.
First reflex: open the replays. Hotjar, Microsoft Clarity, Contentsquare, Mouseflow, FullStory. These tools have become standard in e-commerce, and for good reason: they record every click, every scroll, every hesitation. They're very useful for understanding a journey.
But when the question is "what am I losing right now?", their limits show quickly.
The leak is buried in the mass
Take an established e-commerce site. This month, it recorded 200,000 sessions. Its marketing team watched 40.
| 0.02% | of the site actually seen. Nobody looks at the rest. |
Nobody has time to watch 200,000 sessions, however rigorous the team. So you sample, you follow a hunch, you open the journeys that look suspicious. One session at a time, without knowing which one to look for.
But a replay shows one visitor. A leak shows up across thousands. A promo code reserved for your loyal customers circulating on a deals site? In a replay, a customer enters a code and confirms their order, which looks perfectly normal. Only the volume reveals the leak.
And some leaks leave no trace on screen. An ad pixel cut off after an update: the visitor buys normally, your campaigns stop receiving the conversions. A confirmation email that no longer goes out: everything happens after the click. You could watch every one of your 200,000 sessions and never see them.
Illustrative figures. Run the math with your own volumes: the order of magnitude is often the same.
The error nobody can reproduce
Say you find the right session. The visitor gets stuck while customizing their product, then leaves. You pass it on to the agency. The agency tests it. Everything works.
It's the most common scenario, and the most frustrating. The errors that cost the most are often intermittent: they hit a specific browser, a phone model, a bank, a time of day. A replay shows what happened to one visitor, once. It doesn't tell you how many others are affected, or under what conditions the error comes back.
An online fashion brand went through exactly this: one visitor in ten saw their product customization fail. The agency's manual tests never reproduced the problem. Without proof it exists, it never gets prioritized. So everyone moves on, and the leak continues.
What's missing is the big picture: how many visitors are affected, since when, and how much it costs.
The cost of doing nothing
A leak you can't see doesn't stop on its own. It lasts until someone notices it. And that delay is what drives the bill.
The same leak, with a €63,000 difference
A promo code reserved for your loyal customers ends up on a deals site. Cost: €2,100 of margin per day. Two ways to find out.
Without an alert
| When | What happens | Cumulative loss |
|---|---|---|
| Day 1 | The code starts circulating. | €2,100 |
| Day 3 | You watch a few replays. The orders look normal. | €6,300 |
| Day 15 | The code is picked up by three other sites. Nothing moves in your reports. | €31,500 |
| Day 30 | The monthly review shows margin down. The investigation begins. | €63,000 |
And if nobody makes the connection, the leak carries on into the following month.
With Korvus
| When | What happens | Cumulative loss |
|---|---|---|
| Monday 9 a.m. | The code starts circulating. | €0 |
| Monday 10 a.m. | Slack alert to the e-commerce manager, with the amount. | €88 |
| Monday 11 a.m. | The team deactivates the code. | < €200 |
The leak was the same in both cases. The €63,000 came from the time it took to see it.
Illustrative scenario, calculated on a loss of €2,100 per day.
A French online retailer went through a real version of this scenario: €15,000 of margin lost in one week, because of a promo code used outside its target.
Three habits to put in place this week
You don't need a new tool to start shrinking the blind spot.
- Compare your ad platform sales with your back office every week. If your ad platforms report noticeably fewer conversions than your actual orders, a pixel is probably no longer reporting everything.
- Track how many times your promo codes are used, not just the revenue they bring in. A code meant for a few hundred customers that shows thousands of uses has leaked.
- Place a test order after every release, all the way to the confirmation email. Five minutes, and the only way to check what happens after the click.
These habits have a limit: they're one-off checks. They tell you whether something broke last week, not what is breaking right now.
The problem comes to you
Korvus flips the logic. You no longer go looking for the leak in replays: Korvus monitors every session on your site, continuously, and spots what deviates from normal. It sees the volume, where a replay shows one visitor. It measures how many visitors are affected, where the agency can't reproduce the error.
As soon as an anomaly costs you money, the alert reaches the right person, via Slack, email or SMS, with the amount.
Korvus · Slack · 10:02 Promo code used outside its target. €2,100 of margin per day
Korvus installs in under 5 minutes through a GTM tag, with no work from your tech team, and never asks you to open a dashboard. It doesn't replace your replay tools. It tells you when there's something to understand, and how much it costs.
Your site is losing money right now. Korvus alerts you and tells you how much.