Outbound attribution tracking across multi-step sequences

When a prospect books a meeting after receiving a cold email, two LinkedIn touches, and a call, the question of which touchpoint actually moved them is unanswerable inside any of the current AI sales tools. Attribution is logged at the campaign level — 'this sequence generated 8 meetings' — but not at the touch level. One complaint specifically called out attribution reporting as needing to be 'more detailed,' and others flagged the inability to do per-message-variant analysis, which is the same underlying problem.

This gap exists because multi-touch attribution is genuinely hard to build and requires a unified data model across email, call, LinkedIn, and CRM activity — and no single AI sales assistant owns all of those channels. Each vendor reports on what happens inside their own tool and stops there. Stitching together a cross-channel attribution model requires access to data from multiple systems simultaneously, which creates an integration problem that no individual vendor has strong incentive to solve (it would require them to ingest a competitor's data).

What sales teams currently do: they either ignore attribution entirely and optimize based on gut instinct about what's working, or they have a RevOps analyst manually build attribution models in a spreadsheet by exporting from three different tools and joining on email address. This happens quarterly at best, meaning sequence optimization decisions are made on data that's 90 days stale.

This is a business because every team running multi-channel outbound has this problem permanently. The sequence mix is always changing — new channels get added, new variants get tested — and the question 'what's actually driving meetings' is always being re-asked. The answer determines where SDRs spend their time and where the company buys more tooling, so getting it wrong has real dollar consequences.

What to build

Build a cross-channel attribution engine that ingests email events, call logs, and LinkedIn touch records from the major AI sales assistants via API, joins them by prospect identity, and produces per-touchpoint influence scores and meeting-driver analysis — exportable to CSV and viewable as a filterable dashboard.

Where to start

Start with teams where Salesforce is the shared CRM, use the Salesforce contact ID as the universal join key across all channel data, and pitch specifically to Salesforce-native RevOps teams who already think in terms of contact-level attribution — they'll validate the model fastest and have the clearest ROI story.

The hard part

Prospect identity resolution across channels is the core technical problem — the same person appears as an email address in one tool, a phone number in another, and a LinkedIn URL in a third, and joining these without a CRM as the shared key requires probabilistic matching that will have an error rate that skeptical RevOps buyers will scrutinize closely.

How it makes money

Usage-based pricing tied to the number of active prospects tracked per month, with a low floor for small teams and a meaningful per-thousand rate above 5,000 prospects — aligns cost to the scale of the outbound program.