The waterfall#

Disburse's own data answers first. When a person comes back without a work email or mobile, the waterfall is the next step: a second, explicit ask that looks beyond our data for that one contact, paid from a separate balance of waterfall credits and charged only when something is found.

How it runs#

1. Own data first. Every search, export, lookup and extension reveal is served from disburse's data and costs the usual one credit per lead. None of it spends waterfall credits.
2. The gap is reported. A lookup that found no contact says so, and a search summary counts the rows the waterfall could still fill.
3. You say yes. The waterfall runs only when it is asked for: the waterfall option on enrich_person, bulk_enrich_people, POST /v1/enrich/person and POST /v1/enrich/people, or the Find via the waterfall buttons in the Chrome extension. Agents are told to confirm with you before setting it.
4. It walks the sources in order. Each source is asked only for what is still missing, and the walk stops as soon as the contact is filled. A source that is slow, over its limit or down is skipped and the next one is asked.
5. The employer is checked. A waterfall email is delivered only when the employer it reports agrees with disburse's record for that person. A different employer usually means an old address, so the email is withheld with a note. A mobile follows the person, so it is always delivered.
6. What it finds becomes your data. A contact the waterfall delivered is kept for your organisation and served as disburse data from then on. The next search, export, lookup or reveal that reaches that person carries it, spends no waterfall credits, and the lead itself never costs a credit.

What it costs#

Nothing found means nothing charged. A withheld email is still charged, because the lookup was made and answered; the note tells you why it was held back.

Waterfall credits are a monthly balance separate from your credits, bought on the dashboard's billing page in blocks of 250 up to 10,000 (Growth and Scale plans). They expire at the end of each month and do not roll over. A batch stops at the balance rather than overdrawing it: rows past that point come back with own data only and say so.

How to use it#

In Claude, ChatGPT or any MCP agent. Run a normal search, for example "use disburse and get me 1,000 mechanical engineers in Adelaide". Say the agent brings back 1,000 people and 620 of them carry an email or a mobile. To fill in the rest, ask for the waterfall by name:

Now use waterfall credits to get me mobiles and emails for any contacts that were missing them

The agent runs the waterfall on exactly those rows, charges only what it finds, and reports how many it filled. For a whole list in one go, ask for the export to run with the waterfall and the file comes back with the found contacts in place.

In the Chrome extension. If disburse has no email or mobile on file for the LinkedIn profile you are viewing, press Find email via the waterfall or Find mobile via the waterfall and the contact appears in the panel.

Credits and waterfall credits#

What's the difference between a normal disburse credit and a waterfall credit? A credit buys a lead from our own 88 million strong APAC database. If our database doesn't hold an email or mobile for a person, the waterfall finds it. That spends waterfall credits, a separate monthly balance at 3 cents each: 1 for an email, 10 for a mobile, and only when a contact is actually found. Our own data always answers first and never spends them. A waterfall email is delivered only when it agrees with the person's employer on record, so a stale address is withheld rather than sent to you.

Do I need waterfall credits? Usually not. Our own database covers most of what people ask for. Waterfall credits earn their keep when your target audience is small: an enterprise rep with a handful of accounts to sell into, or someone with a really niche market. They also help when you need the mobile of one specific person. Disburse holds millions of mobiles, but sometimes a seller needs that one number.

Where it is available#

On an export, the rows own data leaves without the contact you asked for go through the waterfall as the file renders, so the file comes back in one piece with the found contacts in place. The start response says how many lookups the current balance covers; rows past that carry own data and the job's waterfall_summary says so. Rendering takes longer than a plain export, so keep polling fetch_export until it reports done. Searches do not run the waterfall; their summary counts the rows it could fill.

Reading the answer#

Every answer that ran the waterfall carries a waterfall block saying what happened per field:

Without waterfall credits#

The ask still succeeds with own data only, and the answer carries a note giving the balance, the cost of the lookup and where to add waterfall credits. Nothing is charged. In the Chrome extension the buttons are replaced by a link to the billing page.

Limits#

The waterfall is paced across all of disburse, so a batch of 100 where most rows need it takes around half a minute. There is a daily ceiling on how much the waterfall can ask; past it, an answer says the waterfall is temporarily unavailable, nothing is charged, and the ask can be repeated later.