Data append

How data append works in CRM enrichment

Data append starts with a record you already have. The CRM holds a person or company key. An append job asks an enrichment provider for missing fields and writes them back. That is the same spine as data enrichment: improve held records, do not pretend the vendor invented your pipeline 1.

Keys gate the append. LinkedIn profile URLs are the usual key for Profile enrichment. Company page URLs are the usual key for Company enrichment. RichAPI prices both at 1 credit per call and bills on any 2xx response, including thin empty bodies 2. Bad keys still cost credits.

Email and phone appends often use waterfalls. RichAPI Email finder, Email verifier, and Phone finder can bill zero when no source returns data 2. That miss policy makes multi-source append cheaper than a single vendor that bills every empty try.

Write rules before the job. Blank-fill only is safest for customer-typed emails. Overwrite may be fine for industry when first-party industry is empty. Mixed rules belong in the runbook, not in a Slack thread after finance sees the receipt.

Score the append with match rate, fill rate, and data accuracy on a frozen sample 3. Coverage without accuracy is just faster wrong CRM data.

Stamp every write with source and retrieval time. Append without provenance becomes anonymous within a quarter and cannot be audited when a field causes a bad send.

Batch append and interactive append share pricing units on RichAPI. The difference is orchestration, pacing, and scorecards across thousands of rows instead of one screen 12.

Idempotency still matters. Re-running the same append should not duplicate activity fields or create conflicting custom properties. Design writes to be safe under retry.

Partial failures need a replay queue. Rows that timed out should not be marked enriched. Otherwise the scorecard lies and the next job skips blanks that never filled.

Coordinate with dedupe. Appending before merge multiplies waste. Merge survivors first, then append once 1.

Tell sellers what changed. A silent industry overwrite mid-quarter creates distrust. A weekly changelog of appended fields keeps adoption high.

Cap concurrency. Bursting append traffic can trip rate limits and produce a mess of retries that look like coverage gaps 2.

Contract the append window. Overnight jobs that rewrite titles during an active sequence create messy personalization. Schedule appends in maintenance windows sellers already expect 1.

Watch for append amplification of bad keys. If a wrong company page URL is stored, every refresh will confidently write the wrong industry. Key repair belongs before credit spend 2.

Keep a short denylist of fields that enrichment must never touch, such as billing email and legal entity name on paid accounts.

Append design choices that change outcomes

Choose fields worth appending. Industry and employee band often lift routing. Random social handles rarely do.

Choose when to overwrite. First-party emails usually win. Stale imported titles may lose to a fresher append when the gold sample agrees.

Choose verification gates. Appending emails without Email verifier at 2 credits per verdict pushes accuracy risk into the sequencer 2.

Choose sample size before full volume. RichAPI match-rate guidance puts match rate next to fill rate and data accuracy for a reason 3.

Choose credit alarms. One-credit profile calls and five-credit email finds add up when append jobs loop on bad keys 2.

Choose refresh cadence. Appended fields still decay. Schedule re-append windows instead of assuming day-one fill lasts a year.

Map append outcomes to ticket types. Wrong phone and wrong industry need different owners.

Keep a rollback plan. If an append source proves bad, you need the prior value. Soft-delete or version fields beat irreversible overwrites.

Align append windows with campaign calendars so sellers are not mid-sequence when firmographics shift under them.

Review append ROI quarterly as credits per verified usable field, not credits per attempted row.

Example of appending industry without overwriting owner email

Ops imports a partner CSV with name, company domain, and empty industry. They resolve company page URLs, then append industry with Company enrichment at 1 credit per call 2.

Blank-fill only is enabled for industry. Partner-supplied domains never change. A sample of a few hundred rows shows fill rate and a spot-check accuracy pass before the rest of the file runs 3.

Rows that still lack industry after the append stay blank rather than receiving placeholder text. The scorecard reports appended fill, credit spend, and leftover blanks.

A second pass appends work emails only for rows marked outbound-ready, using a miss-free waterfall so empty tries do not burn credits 2. Verification follows before enrollment.

That is data append as controlled enrichment, not as a bulk paste from a vendor file.

They store before-and-after values for industry on the sample so overwrite bugs are visible. No unexpected clobber of partner notes appears.

The next partner sync reuses the same blank-fill map and credit alarms. Append stops being a heroic script and becomes a boring weekly job 23.

Ops files the credit receipt beside the fill and accuracy numbers so the next partner sync starts from evidence, not from memory of a good week.

Data append vs data enrichment

Data append is the field-writing move. Data enrichment is the broader practice of enhancing held records, which usually includes append plus verification and refresh.