Conversations-booked attribution when email and LinkedIn both touched the prospect

Last-touch reporting will lie to you the moment email and LinkedIn share a prospect. The meeting booked after a LinkedIn DM often started with an email that earned the connection. Credit only the DM and you will overfund LinkedIn steps and underfund the opener that made the DM possible.

GTM runs email and LinkedIn inside the same client-isolated workspace and optimizes toward booked conversations, not vanity reply rate. Attribution is how you keep that optimization honest when both channels touched the same person.

What you are actually attributing

For agency outbound, the conversion event that matters first is conversation booked (qualified meeting accepted on calendar), not open rate and not raw reply count. Pipeline and revenue come later. If your booking event is fuzzy, every model downstream is theater.

Define the event once per client:

• What counts as booked (Calendly accept, AE-confirmed meeting, etc.)
• Which contacts can receive credit on an account
• Lookback window from booking backward
• Whether influencers on the buying committee can share credit

Apollo-style outbound attribution guidance keeps pushing the same backbone: log activities against contact and opportunity, then separate sourced pipeline from influenced pipeline. Agencies should do the same per client workspace.

Why single-touch breaks multichannel outbound

Typical messy path:

1. Day 1 email opener
2. Day 3 LinkedIn connection accept
3. Day 5 LinkedIn note
4. Day 8 email bump
5. Day 9 calendar book from the email reply

Last-touch says "email bump won." First-touch says "email opener won." Both erase the LinkedIn acceptance that made the bump land differently. Linear or position-based models exist because outbound paths are short enough to list and too entangled for one hero touch.

A practical model agencies can defend

Start simple. You can get fancy later.

Recommended default for agency client reporting:

• Log every touch with timestamp, channel, campaign/sequence ID, contact, account, and outcome.
• Report pattern credits alongside a primary model:
• email-only bookings
• LinkedIn-only bookings
• email -> LinkedIn bookings
• LinkedIn -> email bookings
• mixed with call (if you place calls)
• Use position-based or linear credit for the primary internal model on mixed paths.
• Keep first meaningful touch as a separate column for list and offer quality.
• Keep last human reply before book as a separate column for handoff quality.

Reachly and Cometly-style outbound attribution writeups converge on the same inputs: touch timestamps, channel IDs, prospect IDs that stitch channels, and a clean outcome event. Without those four, the model is a slide, not a system.

Instrumentation checklist per client workspace

1. Same person identity across email and LinkedIn (or an explicit merge rule).
2. Auto-logged sends, accepts, replies, and booking events into one timeline.
3. Sequence stop-on-reply so channels do not keep touching after the human thread starts.
4. Booking objects tied to the contact and, when relevant, the company.
5. No double-counting when both an email reply and a LinkedIn reply exist for the same booking.
6. Weekly export the client can read: bookings by pattern, not only by last channel.

If LinkedIn activity only lives in a rep's native inbox, attribution will keep missing half the path. The workspace has to see both channels.

How to use the numbers without politics

• Do not cut email because last-touch favors LinkedIn for a month.
• Do not cut LinkedIn because first-touch favors email.
• Change step design when a pattern repeatedly books (for example email opener + LinkedIn accept + short DM) and another pattern repeatedly stalls.
• Show clients pattern charts. They understand "both channels showed up" faster than a fractional credit debate.

Dievio's multichannel attribution framing is useful here: pick linear or time-decay when you need a single score, but keep channel-combination reporting so operators can act.

Guardrails that keep agencies out of trouble

• No invented lift percentages in client decks.
• No competitor benchmarks presented as your client's truth.
• No blending two clients' attribution in one rollup without isolation.
• Reconcile booked conversations to CRM opportunities on a schedule, or stop calling it pipeline attribution.

Where GTM fits

Find -> qualify -> write -> send/reply/book is one motion across email and LinkedIn in an isolated workspace. Attribution is the scoreboard for that motion. If the scoreboard only celebrates the last click, ops will optimize the last click and slowly delete the touches that made the meeting possible.

Credit the path. Report the pattern. Keep the booking event clean. That is enough to steer an agency outbound program without pretending you solved marketing attribution science.

Sources: Reachly multi-touch attribution for B2B outbound, Cometly multi-touch outbound attribution, Dievio outbound multichannel attribution, Apollo outbound pipeline attribution, GTM product knowledge.