Start with the arithmetic, not the channel
Before a campaign exists there are four numbers: average order or deal value, gross margin after all variable costs, the conversion rate from lead to sale, and the contribution the business needs to keep. Those four decide the maximum acceptable cost per acquisition, and therefore whether a channel is viable at all.
Most 'the ads stopped working' conversations are actually a margin conversation that was never held. The break-even ROAS calculator settles it in about ninety seconds.
Funnel architecture
Match the offer to intent
Search intent buys solutions; social intent buys interruption. The same landing page cannot serve both, and a lower cost per lead from a mismatched audience is a more expensive customer later.
Design the conversion step around the next human action
The form exists to start a conversation the team can actually hold. Ask for what routing and qualification need — nothing more, because every extra field costs volume, and nothing less, because missing context costs response time.
Instrument before launch
Tracking, event definitions, source integrity and the CRM record path are QA items, not follow-ups. Launching untracked spend buys data you cannot use.
Close the loop to revenue
Optimise against the outcome the business sells, not the platform's proxy for it. Where the closed-loop path does not exist yet, build it before scaling budget.
What I actually do inside accounts
- Structure campaigns so that budget decisions can be made per intent, not per platform habit.
- Set measurement first: conversion definitions, event hygiene, and the boundary of what the platform can claim.
- Test at the level that changes outcomes — offer and audience before creative variants, creative before bid tinkering.
- Watch cost per qualified outcome, not cost per lead. Cheap leads that never qualify are the most common way to lose money slowly.
- Keep a change log so performance shifts have a documented cause.
The measurement conversation
| Metric | Good for | Cannot tell you |
|---|---|---|
| Platform ROAS | In-platform optimisation and pacing | Whether the business made money |
| Cost per lead | Channel efficiency at the top | Whether the lead was qualified or reachable |
| Cost per qualified opportunity | Commercially honest channel comparison | Delivery cost and retention |
| CRM-sourced revenue | Budget allocation | The full assist path, which is genuinely unknowable |
The published numbers on this site follow the same discipline — see the case studies, where each figure carries its own scope and its own limits. The messaging campaign record is the clearest example of why acquisition volume is an operations question: thirty-three thousand conversations is a staffing decision before it is a marketing result.
E-commerce specifics
- Margin varies by product; a blended ROAS target hides the products losing money.
- Returns and payment fees belong in the margin calculation before setting targets, not in a quarterly surprise.
- Feed quality and product data usually beat bid strategy for shopping performance.
- Post-purchase operations — fulfilment accuracy, delivery communication, support response — determine repeat rate, which is where the actual profit lives.
Where this connects
Acquisition feeds the lifecycle described in marketing operations; the failure it exposes most often is leads going cold after the form. If growth is already outpacing delivery, the more urgent page is scaling breaks operations.
