Skip to main content

Integration phases

Bringing your data into DQC is a multi-step engagement, not a single switch. We run it in three phases so the mapping firms up against real data before anything goes live.

Figure: the three phases. Phase 2 loops back while we agree what your values mean.

Phase 1 — Data collection​

You start sending. Each respondent is POSTed to DQC and we store exactly what you agreed to send.

Nothing is mapped yet, and that is deliberate. We collect first so that when we write the rules we are writing them against records that actually arrived, not against a spec of what was supposed to arrive. This is the phase to agree on now, and it is the only phase that needs work from you.

What you need to have settled before you start:

  1. ✅ Your DQC API key.
  2. ✅ Which fields you can send — see POST /data.
  3. ✅ A stable id per respondent, so two sends about the same person join up.

Phase 2 — Mapping​

Once data is flowing, DQC writes and iterates the mapping rules against what we actually receive. We analyse what arrives, build the rules, and refine them as your variables firm up.

This is hands-on, and it is the phase that takes real time — realistically a few weeks, depending on your integration and how many fields you send.

Set expectations here

Not every field maps automatically. Some values need discussion: we may hit exceptions, or find we are not receiving a value we expected, and we will come back to you with clarifying questions.

That back-and-forth is normal. It is exactly why we collect data first and map second.

You do not have to reshape anything to help this along. Your status codes, your termination reasons, your quality-check names — send them in whatever form you already use and we write the translation. See Sending Data to DQC.

Phase 3 — Live​

Stored data is mapped and written on an hourly cycle, and the dashboard updates every hour.

Nothing changes on your side between Phase 1 and Phase 3 — you are already sending the right thing. Phase 3 is us turning the mapping cycle on.

Near-real-time, not live per response

Ingestion is fire-and-forget in every phase: you send, we return 200, your respondent moves on. Phase 3 adds the hourly mapping cycle on top of that. A record will not appear in the dashboard the second you send it, and that is expected.

✅ Summary​

  1. Phase 1 — you send, we store. The only phase that needs engineering work from you.
  2. Phase 2 — we map. A few weeks, with clarifying questions along the way. Normal, and the reason we collect before mapping.
  3. Phase 3 — live. Hourly cycle, dashboard updates every hour. No change on your side.
  4. Ingestion is fire-and-forget throughout. 200 means stored, not mapped.