Home / Notes

Rebuild or optimize? When an account is past saving

Optimize when the account's structure and measurement are sound and only settings drift; rebuild when conversion tracking is untrusted, campaigns overlap and fight each other, or years of patches have made cause and effect unreadable. The test: if you can't say what any single change did, the account is telling you it needs a rebuild.

Every inherited account arrives with the same question: can this be fixed in place, or does it need to come down? Agencies default to "optimize" because rebuilds are scary to sell. Owners default to "optimize" because it sounds cheaper. Both are often wrong, and the account itself will tell you — if you know which five signals to read.

Five signs the account is past optimization

One: conversion actions nobody trusts — duplicates, legacy imports, form-fills counted as revenue events. Two: campaign overlap, where the same query can be served by three campaigns and budget flows to whichever bids worst. Three: bidding strategies fighting structure — tROAS on campaigns whose conversion values are fiction. Four: a change history so dense that no single change can be attributed an outcome. Five: performance that ignores your inputs — you cut budgets, pause keywords, and the numbers barely move, because automation is routing around a broken signal.

Why patching a broken structure fails

Smart Bidding learns from the conversion signal it's fed. If that signal is wrong — junk form fills, duplicate counting, no CRM truth — then every optimization you layer on top teaches the machine to be wrong faster. This is why accounts sometimes get worse after diligent weekly optimization: the work is real, the direction is wrong. Fixing the signal inside a tangled structure is often more surgery than replacing the structure around a clean signal.

What a rebuild actually changes

A rebuild is not "new campaigns." The sequence is: verify measurement against the CRM first — what counts as a conversion, imported from where, deduplicated how — then design the fewest campaigns that give Google clean signals per business outcome, then a negatives architecture that enforces the boundaries, then bidding chosen to match actual conversion volume. Structure serves signal, not the other way around. I don't restructure on top of broken tracking — that just makes the wrong number go up faster.

Migrating without torching performance

The fear is the learning reset. It's manageable: keep the old campaigns running while the new structure launches on a budget slice; protect the pockets of genuine learning (usually brand and one or two proven exact-match themes) by migrating them last; move budget as the new structure proves itself against CRM outcomes, not platform CPA. Short turbulence can happen. Flying blind forever is worse.

The honest cost comparison

Optimization-in-place looks cheaper monthly and often costs more annually: you pay the retainer and the waste the structure protects. A rebuild is a bounded project with a before/after you can verify in your own CRM. The deciding question isn't the fee — it's whether another six months of the current account is an asset or a liability.

VERIFIED EXAMPLE · When I inherited a wasteful account at a luxury US healthcare clinic, it was optimizing toward platform conversions. Rebuilt around admitted patients imported from the CRM, cost per admitted patient fell ~78% while spend scaled ~2× and approved patients rose ~9×.

Five signs the account is past optimization
SignWhat it looks like
Untrusted conversionsDuplicates, legacy imports, form-fills counted as revenue
Campaign overlapThe same query served by three campaigns at once
Bidding fighting structuretROAS running on conversion values that are fiction
Dense, unreadable change historyNo single change can be credited with an outcome
Inputs ignoredBudget cuts and pauses barely move performance

Questions owners ask

How do I know if my tracking is actually broken?

Reconcile one month: conversions Google Ads claims vs revenue-stage records in your CRM for the same period. A 10–20% gap is normal. Two to five times is a broken setup steering your bidding.

Will a rebuild reset all my account history?

Account-level history and quality signals persist. Campaign-level learning resets where campaigns are replaced — which is why migration is sequenced and proven pockets are protected rather than everything moving at once.

Can my in-house person execute the rebuild plan?

Often yes. A rebuild plan with the new architecture and migration map is a deliverable your team can execute — the difference is whether you want the design or the design plus the hands.

Not sure which side of the line your account is on?

Thirty minutes with your account on screen. I'll tell you whether it's an optimize, a rebuild, or a tracking fix first — and I'll say so if it's fine as it is.

$500 audit + 90-day roadmap · direct with me, no account managers.

Book a 30-minute call