Analytics
Why did it do that?
That is the only analytics question that matters when a customer complains, and most tools cannot answer it. EzyConn keeps a timeline for every process run: which step ran, what the customer answered, which rule decided what, and how it ended. Six weeks later you can open it and read the whole thing back.
Why it matters
A dashboard of conversation counts tells you the tool was busy. It does not tell you whether the customer who rang to complain was turned away correctly, or which question forty people abandoned last week.
Both of those are answerable, but only if the run was recorded properly at the time. Once it has been rolled up into a monthly number, the answer is gone.
So the record comes first here and the summary second. The counts are built out of runs you can open, rather than the runs being thrown away once they have been counted.
How it works
What one run record contains.
Four things, kept for every process run, whichever channel it arrived on.
- Which step ran
- In order, from the question that opened it to the ending it reached. You can see where a customer stopped answering, which is usually the question that needs rewording.
- What the customer answered
- The values the process collected — the service, the postcode, the urgency, the contact details — as they were given, attached to the run they came from.
- Which rule decided what
- The deterministic checks and their results. Do we cover this postcode? Is this a number that actually dials? Is this time still free? Each one is recorded with the answer it produced, so an outcome can be explained rather than guessed at.
- How it ended
- The named outcome the run reached, from the list that process declared it could reach. Nothing ends as "closed" with no further detail.
The endings
11 ways a run can finish, and it is always one of them.
Every published process declares the endings it can reach before it ever runs. That is what makes the reporting countable: a result is a named outcome, not an interpretation of a transcript.
- Appointment booked
- Appointment moved
- Appointment cancelled
- Outside the area you cover
- Nothing free — handed to a person
- Handed to a person
- Enquiry captured and routed
- Visitor left
- Something went wrong — handed to a person
- No matching appointment — handed to a person
- More than one match — handed to a person
Who uses it
Three people, wanting three different things.
All 24 published processes record the same way, so the answer does not depend on which industry you are in. Browse the catalogue.
The owner, on a Monday
How many appointments came out of it last week, and how many enquiries went to a person instead.
Whoever is fixing the wording
The step people abandon, the rule that turns the most customers away, and whether the new version of a question did better than the old one.
Whoever takes the complaint
The single run in question, opened, with what was asked, what was answered and which rule produced the result.
What it can do
Six things, none of them a percentage.
Open any past run
History is kept per run, not summarised away. A conversation from six weeks ago opens with its whole timeline intact.
Count what happened, by period
Conversations, bookings and outcomes over a period you choose, so you can see whether last month was actually busier or just felt it.
Compare two wordings
Variants let you run two versions of a process and see which one reached the outcome more often, instead of arguing about which question sounds friendlier.
See where people leave
Runs that ended without a result show the step they got to. Three in a row stopping at the same question is a wording problem you can fix in an afternoon.
Watch the endings, not the chatter
Because every process declares its outcomes up front, the report is a count of real results — booked, moved, cancelled, routed, handed over — rather than message volume.
Know what you are using
Messages and AI credits are metered per period on every tier, including the free one, so usage is a number you can plan against rather than a surprise.
What it works with
Sixteen real events leave EzyConn as things happen — bookings created, confirmed, moved, cancelled, completed or marked a no-show, conversations opened and closed, contacts created — through Zapier or a signed webhook to your own endpoint.
That is the route into your own spreadsheet, warehouse or reporting tool. It is a stream of events rather than a scheduled export, so what you build on it is as current as the last thing that happened.
There is no Google Analytics connector — it is declared in the product’s own catalogue and not built — and no CRM or helpdesk integration to push figures into.
What plan includes it
Analytics and reports are on every plan.
Seeing what happened is not something we hold back for a higher tier. What sits above it is the interpretation layer, and your own usage meter is visible everywhere.
- Run history, outcomes and per-period analytics
- Free
- Reports
- Free
- AI Insights
- Professional
- Your own usage meterEvery tier is metered — the free one at 500 messages and 1,000 AI credits per period.
- Free
Read this before you buy
What analytics cannot do.
The limit: EzyConn does not report a resolution rate, a deflection rate, a satisfaction score or a return-on-investment figure — not on this website and not in your account. Every one of those numbers depends on judging whether a customer was actually helped, and nothing here can honestly make that judgement. Any vendor quoting you one has decided what counts as resolved on your behalf.
It only sees what EzyConn did
A customer who gave up and rang you, or emailed an address EzyConn does not read, is invisible here. Nothing from your other systems is counted.
No industry benchmark
There is no “businesses like yours” comparison, because we have no dataset that would make one honest. We also publish no traction figures, customer counts or case-study numbers anywhere on this site.
No customer satisfaction score
EzyConn does not ask your customers to rate anything, so there is no score to report. If you want one, send an event out through Zapier and use a survey tool that does it properly.
The record is of the process, not the person
Free-text answers are stored as the customer wrote them, in the appointment notes, with no retention rule and no consent record. Do not use free-text questions to gather anything sensitive.
No data-residency or export guarantee
There is no choice of region, no self-hosted option and no scheduled bulk export. Getting data out means the event stream, through Zapier or your own webhook endpoint.
No uptime or retention commitment
EzyConn publishes no service-level agreement, and no promise about how long run history is kept. If a record has to exist in five years, keep your own copy of the events.
Judge it on the runs, not on a dashboard.
Open ten conversations and read what actually happened in them. That is a better measure of whether this works for you than any figure we could print.