Your analytics show the drop and stop there. hedlyte explains that drop — the reason those shoppers gave up, what losing them costs you every month, and the one change that wins them back. Then it goes hunting for the next leak.
A brief, not a dashboard — a new one each time there's money to win back.
You're losing $4,200/mo at checkout.
Phone shoppers add to cart — then leave the moment shipping appears at the final step. A surprise cost, right before they pay.
Real output from hedlyte running on a live store. Yours is computed from your own shoppers.
You pay for the traffic. Most of those shoppers leave. You never find out why.
Two in three visitors you paid to bring in leave without buying. Your analytics can tell you how many, and which step — never the reason. So the number gets reviewed, nothing gets decided, and the leak keeps running quietly on every order you don't get.
The kind of thing hiding in a checkout right now — invisible on every chart you own, obvious the moment something actually reads the behaviour behind the number. This is what lands in your inbox.
Every frame below is real output from a live store, shown unnamed — so you know exactly what arrives in yours.
$3,100/mo
About 1 in 9 attempts is declined with no reason shown. The shopper sees "try again," tries once, and gives up.
412 shoppers/mo
Shoppers click your campaign link, paste the code at checkout, and it comes back "invalid." They don't ask why — they close the tab.
$1,900/mo
The size the shopper added is gone by the time they reach payment. No alternative offered, no notify-me — just a dead end.
700 shoppers/mo
Shoppers tap the size they want and nothing happens. They tap again, then leave without adding to cart.
Free to watch · a few minutes · the findings unlock with an email
Connect once, on any platform. From then on it works the way a good conversion team would — continuously, in the background, without being asked. The difference is the last step: it comes back to check whether it was right.
One async snippet, on any platform. It never blocks a page.
It reads the sessions and finds where the money leaks — the step, the segment, the monthly cost.
You get the leak, who it hits, the change to make, and the clips that prove it.
Ship the fix and it measures the same shoppers again — says whether it worked, and remembers either way.
Every answer makes the next one sharper. That loop is the part nobody else runs.
It doesn't guess and it doesn't answer from a summary. It measures first, narrows down to who is actually affected, then reads the recordings of those exact shoppers — and shows you every step it took to get there.
Mobile Safari shoppers are 4× more likely to abandon at payment.
I read 12 of them. In 9, the shopper tapped Pay, the keyboard opened over the button, and nothing happened again for 40 seconds or more. Desktop finishes at 3.1%; this group finishes at 0.7%.
Every line above is copied from one real run, including the timings.
Every finding is specific enough to act on today — never a vague "users are confused." And when your funnel's healthy, hedlyte says so. No manufactured problems.
What a finding actually reads likeFROM A REAL STORE"Mid-range Android shoppers can't see the Pay button — the keyboard covers it. 214 of them dropped here last week. Fix: move the button above the fold."
Charts, funnels, segments, exports — more detail every year. Those tools tell you a number moved. Turning that into a reason, and a reason into a decision, is still your job, and it still takes hours you don't have.
It reads the behaviour behind the number, works out the reason, prices it, and pushes you the decision with the fix attached and the proof underneath. You read a memo, not a dashboard.
We're onboarding stores in batches. One snippet, on any platform, and hedlyte starts reading real shoppers — your first answer lands as soon as there's enough to be sure of, and a new one whenever there's money to recover.
Scoped to the funnels you choose · your data stays yours · works on any platform
Most online funnels are standardized — product page, add to cart, cart, checkout, purchase — and most of the traffic is mobile. That's exactly why the leaks are easy to miss: a checkout that works perfectly on your desktop can be quietly broken on a mid-range Android, and nothing in the totals would tell you. Cart abandonment and checkout drop-off show up as a percentage on a chart with no explanation attached, and a percentage has never once told anyone what to change.
The highest-ROI problems are also the most neglected, because none of them are visible in aggregate: failed payments, invalid-coupon errors, a shipping cost that only appears at the final step, an out-of-stock surprise after checkout has started. Each one shows a message to a real shopper and sends them away. Each one is invisible until something reads what actually happened, at the level of the individual journey, and counts how often it repeats.
Not every drop is a defect. When a leak concentrates in one segment — a single device, browser, or step — it's usually something broken that you can go fix this week. When it's spread evenly across everyone, it isn't a UI problem at all; it's the offer, the price, or the copy. Telling those two apart is the difference between shipping a fix that moves the number and burning a sprint on the wrong thing entirely.
That judgment is what hedlyte automates. It examines the shoppers who dropped, separates what's certain from what's merely worth testing, prices the gap in the unit that matters to you, and gives you the cheapest experiment to run next. It's the analysis a conversion specialist would do by hand — running continuously in the background, on any platform, and reporting back only when there's something worth your attention.