Skip to content
NewAI dashboards.See how →
Open MammothTry free
Workflow · Search and ads · Opening soon

Multi-touch attribution

Split each closed deal across the touchpoints before it. You can see first touch, last touch, linear, and channel views from HubSpot exports.

Early access opens shortly. Check back soon.

Book a demo
Someone typed

“show each closed hubspot deal by pre-close touch, split credit evenly across touches from the deals, touchpoints, contacts, and channel mapping files; Status says "Agrees" when the touch is mapped and before close, else what's off”

The agent read it as

show each closed hubspot deal by pre-close touch, split credit evenly across touches from the HubSpot deals export, HubSpot engagement touchpoints export, HubSpot contacts export, and Channel mapping sheet files; Status says "Agrees" when the touch is mapped and before close, else what's off

What the agent built

8 steps, each one an ordinary Mammoth step you can open and change.

  1. File
    HubSpot deals export
    CSV export · 12 rows
  2. File
    HubSpot engagement touchpoints export
    CSV export · 30 rows
  3. File
    HubSpot contacts export
    CSV export · 12 rows
  4. File
    Channel mapping sheet
    CSV export · 7 rows
  5. Step
    Filter closed deals
    Step 5
  6. Step
    Shape touch ids
    Step 2
  7. Step
    Shape contacts
    Step 1
  8. Step
    Shape channel map
    Step 4
  9. Join
    Add contact details
    30 of 30 paired
  10. Join
    Bring touches onto deals
    12 of 12 paired
  11. Join
    Add channel mapping
    30 of 30 paired
  12. Step
    Split credit and flag
    Step 8
  13. Result
    Closed deal touch credit
    30 rows · 1 flagged
  14. Schedule
    Runs when you say
    Proposed

Checked on the sample files: every step ran, 30 rows came out, and the join paired 30 of 30 rows; the join paired 12 of 12 rows; the join paired 30 of 30 rows.

What you get back

The result of that run, the rows to look at first.

Closed deal touch creditSample data30 rows
29 agrees1 after close
Deal keyDeal nameClose dateDeal amountTouch IDTouch dateTouch contact keyContact nameContact emailTouch channelChannel groupTouch subjectStatusCredit amount
d_1004Harborview Q4 Upsell2026-10-1822,100e_30292026-10-28c_2006Sofia Lopezsofia.lopez@harborview.exampleemailowned_mediaLate follow-up after closeAfter close0
d_1001Northstar Platform Rollout2026-10-2718,400e_30012026-09-10c_2001Talia Brookstalia.brooks@northstar.exampleemailowned_mediaIntro sequence openedAgrees4,600
d_1001Northstar Platform Rollout2026-10-2718,400e_30022026-09-14c_2001Talia Brookstalia.brooks@northstar.examplewebsiteowned_mediaPricing page sessionAgrees4,600
d_1001Northstar Platform Rollout2026-10-2718,400e_30032026-10-03c_2002Jonah Reedjonah.reed@northstar.examplephonesales_outreachDiscovery callAgrees4,600
d_1001Northstar Platform Rollout2026-10-2718,400e_30282026-10-20c_2002Jonah Reedjonah.reed@northstar.examplewebsiteowned_mediaDuplicate session loggedAgrees4,600
d_1002Aster Labs Expansion2026-10-2412,650e_30042026-09-19c_2003Iris Nguyeniris.nguyen@aster.examplesocialpaid_or_socialConnection requestAgrees6,325
d_1002Aster Labs Expansion2026-10-2412,650e_30052026-10-08c_2003Iris Nguyeniris.nguyen@aster.exampleemailowned_mediaProposal follow-upAgrees6,325
d_1003Bluepeak Security Renewal2026-10-219,600e_30062026-09-07c_2004Evan Shahevan.shah@bluepeak.exampleeventearned_mediaSecurity webinar attendedAgrees3,200
d_1003Bluepeak Security Renewal2026-10-219,600e_30072026-09-23c_2005Marin Foxmarin.fox@bluepeak.exampledirect_mailofflinePostcard dropAgrees3,200
d_1003Bluepeak Security Renewal2026-10-219,600e_30082026-10-09c_2004Evan Shahevan.shah@bluepeak.examplephonesales_outreachProcurement check-inAgrees3,200
d_1004Harborview Q4 Upsell2026-10-1822,100e_30092026-09-25c_2006Sofia Lopezsofia.lopez@harborview.exampleemailowned_mediaDemo inviteAgrees7,366.67
d_1004Harborview Q4 Upsell2026-10-1822,100e_30102026-10-01c_2006Sofia Lopezsofia.lopez@harborview.examplewebsiteowned_mediaCase study pageAgrees7,366.67
30 rows in the result · 12 shown, problems first. Sample data: illustrative figures, not business results.

How do you do multi touch attribution?

Export your closed deals, contacts, and engagement touchpoints from HubSpot, then add a channel mapping sheet. The workflow matches touches to deals, keeps only touches before close, and splits credit across the touches. It also flags each row, so you can see which touches agree with the rules and which ones do not.

How it works

How to use the Closed deal touch credit workflow

  1. 1Get early accessWe invite people in small batches, job by job. You'll land on this workflow, already built and checked on the sample files.
  2. 2Swap in your exportsUpload your own files as they come out of your tools. The agent checks the workflow on your rows.
  3. 3Change it in words, then turn it onSay what to change and see the new result before anything saves. Then choose when it runs.

Closed deal touch credit: questions

What export should you use?
Use the HubSpot deals export for closed deals, the HubSpot engagement touchpoints export for the touch history, and the HubSpot contacts export for contact details. Add a channel mapping sheet so each touch can be grouped by channel.
What happens to rows that do not match?
Rows that do not match the rules stay in the output with a status flag. For this workflow, a row can show that the touch is not mapped, or that it happened after the deal closed.
How do you change the attribution rules?
Change the rules in words by editing the step that splits credit. You can switch the weighting rule, change the date logic, or update the channel mapping, then rerun the workflow.

Get early access to this workflow

It will open already built and checked on the sample files. Swap in your own exports and see your result.

Early access opens shortly. Check back soon.

Book a demo

Workflows like this

Decide which pages to keep, update or prune, and find the search terms that spend without converting.

All search and ads templates →

Keep reading

Read nextAll workflowsDescribe yours, or start from one the agent already built.Take a lookProductData EngineConnected sources, cleaning and scheduled refresh.Take a lookConnectorsConnectorsBring the data in from the tools you already use.Take a look