Every tool holds a piece of the same customer. None of them know each other.
Your helpdesk knows they complained. Your email tool knows they never opened it. Your analytics knows they came back twice last week. Your product knows their payment failed. Four systems, four fragments, one person — and nobody puts them together.
Feedbank is where they meet. It matches every source onto one record per person, then tells you, in a sentence, when someone matters enough to do something about.
They each know a piece of the same person
Connect what you already run. Feedbank does the matching — you do not change a single tool.
Zendesk
Intercom
Google Analytics
Your product API
Email marketing
Your website
BigQuery
CSV
/ WHAT FEEDBANK ACTUALLY DOES
Collect. Match. Tell you.
One record per person, not per tool
The same customer arrives as an email address in one system, a user id in another and an anonymous browser in a third. Feedbank matches them onto a single record, so their history reads as one story instead of four disconnected logs.
The whole timeline, in order
Read the pricing page twice. Opened a ticket. Ignored three emails. Payment failed on Thursday. No single tool can show you that sequence, because no single tool saw all of it.
Told about the ones that matter
Set an alert on the people and events worth interrupting you for — a named customer, a failed payment, a complaint from someone who was about to renew. Everything else waits until you look.
[ 01 / 04 ]
RECORDS
One person, however many tools
- Matched on email, user id or browser
- Existing records updated, never duplicated
- Anonymous history attaches on sign-up
- Companies rolled up from their people
Every source updates the same record rather than creating another copy. Match on email, on your own user id, or on a browser that later signs in — the history follows the person, not the tool that saw them.
See a record[ 02 / 04 ]
THE TIMELINE
What happened, across everything
- Every source on one timeline
- Support, product, marketing, web
- The evidence behind every finding
- Readable without training
Support tickets beside page views beside failed payments beside unopened emails — in the order they happened, for one person. This is the view no tool in your stack can produce on its own.
See the feed[ 03 / 04 ]
FINDINGS
Sentences, not dashboards
- Customers who have gone quiet
- People who tried to reach you and could not
- Repeat interest before a purchase
- Friction that costs you people
A rule engine reads the joined-up behaviour and writes what it notices — with its evidence, so you can check it rather than take it on faith. It stays silent when there is nothing honest to say.
How it decides[ 04 / 04 ]
ALERTS
Told, not asked to look
- Subscribe to any kind of finding
- Watch a single named person
- Email and in-app
- Routed to whoever owns it
Choose what is worth an email and what can wait. Watch one named customer and hear every time they come back. Rate-limited on purpose, so a busy day never becomes an inbox to clear.
Set up an alertConnected in an afternoon. Joined up from the first import.
01
Connect what you already run
A tag for your website, an API for your product, a key for your helpdesk, a file if that is simpler. Nothing changes inside the tools themselves.
02
Watch the records merge
Feedbank matches each arrival to a person it already knows, or creates one. The same customer stops being four rows in four systems.
03
Let it read a week
Findings need enough behaviour to be more than a coincidence. Feedbank waits rather than guessing, and tells you exactly what it is waiting for.
04
Choose what to be told
Subscribe to the findings worth interrupting you for. From then on you are told, and you stop having to remember to look.
/ HOW IT BEHAVES
Built to be trusted rather than impressive.
0
However many tools describe them. That is the whole idea — everything else follows from it.
0min
Findings are recalculated every five minutes, so you are told what is happening rather than what happened yesterday.
0
When there is nothing worth saying, Feedbank says nothing and explains why. That is the only reason to believe it when it speaks.
Reasonable questions
Start freeNo. A CRM is where your team records what it did. Feedbank is where your systems record what the customer did, matched onto one person. There is no pipeline, no deal stages and nothing to keep up to date — it reads what already happens and tells you when it matters. It sits alongside your CRM.
By email, by the user id your own product already uses, or by a browser that later signs in and gives itself a name. Where a source offers none of those, that data stays attached to the source rather than being guessed onto somebody — a wrong match is worse than no match.
No. Everything stays where it is and keeps working the way it does. Feedbank reads from them; it never asks them to change, and nothing depends on it being there.
Analytics tells you 62% bounced. Feedbank tells you that the customer who opened a ticket on Tuesday had already read your pricing page twice and has not opened an email since March. One is a measurement of a crowd; the other is a person you can ring.
No advertising cookies and no third-party tracking. Anyone with Do Not Track or Global Privacy Control switched on is not measured at all. Each business's data is held separately and enforced by the database rather than by the application remembering to ask, and it can be exported or deleted whenever you say.