# Do we still need Segment if Croct builds profiles and audiences itself?

Asked by Colin Farrelly on 2024-07-23. Tags: segment, cdp, architecture.

Honest question from someone drowning in martech line items. We pay for Segment
and if I am truthful with myself, 80% of what we use it for is building
audiences that get synced back to the website... with enough delay that the
visitor who triggered the audience has usually left by the time they qualify.

Now we are adopting Croct, which apparently builds profiles and audiences
itself. So before renewal season: what actually overlaps, and what does Segment
still do that Croct does not? Trying to figure out if I can retire a line item
or if I am about to make a classic ops mistake

## 2 answers

### Accepted answer from Croct Bot (2024-07-24)

Croct does not need an external CDP. Profiles, [identity resolution](https://docs.croct.com/explanation/user-profiles/identity-resolution), and
audiences are built in, and audiences are CQL conditions evaluated in real
time on each interaction rather than batch-computed and synced. So the
specific loop you describe, building an audience elsewhere and syncing it
back to the site with a delay, is exactly the part Croct replaces. The
visitor qualifies during the visit where the behavior happens.

What does not overlap: Segment's core job is routing event data to many
downstream tools, warehouses, email platforms, ad destinations, and so on.
Croct acts on data on-site; it is not a general-purpose event router. If
other teams depend on Segment destinations, that need does not go away.

For getting profile data out of Croct into other systems, there is an export
API and CSV export, which covers warehouse-bound use cases but is not a
replacement for a fan-out pipeline.

So the question for renewal season is really: how much of your Segment usage
is the audience-sync loop versus destination routing?

### Answer from tom_b (2024-07-25)

Been through this evaluation. One nuance to add: check who owns the tracking
plan before you cut anything. Our Segment schema had become the de facto
contract between product and data teams, and even after we moved the
audience use case to Croct we kept Segment for a transition period purely as
the routing layer while the data team rebuilt their pipelines. It ended up
being a scope reduction rather than a hard cut, which was much easier to get
signed off.