The average looked healthy. Several locations were invisible.
Thirty trajectories nobody can hold at once
An agency account manager carrying several clients across dozens of locations cannot mentally track every one. They track the clients who contact them most. This is a completely different set from the clients who need attention most.
The reporting usually reinforces this. An estate or portfolio average is stable, easy to produce and comfortable to present, and it hides exactly the locations that are in trouble.
Locations had been checked, but never the same way twice
Client locations had been spot-checked over time by different people, from different places, on different keywords. Assembled into a monthly figure, the result described the measurement as much as the performance.
Measured properly on one common frame, the range between the strongest and weakest location was very wide. Several sites were not appearing at all for the terms that mattered to them. And had not been for months.
Nothing in the previous reporting had shown this, because the average stayed flat throughout. That is the specific failure mode a portfolio average produces, and it shows at renewal.
Make the locations comparable, then work the outliers
The first task was not optimisation. It was making a comparison possible at all.
Put every location on one frame
Same grid geometry, spacing and keyword set for every client location. So the ranking between them reflected performance rather than measurement.
State the method in the report
A ranking without its method invites the reader to assume comparability. Stating it is what lets a client check the figures.
Group before ranking
Locations grouped by market and maturity. Because a mid position in a dense city can be a better result than a top position in a thin one.
Fix the outliers in bulk
The same category and service-list corrections applied across the affected locations at once, rather than one profile at a time over several weeks.
The invisible locations were the fastest to move
The locations that had been failing longest were also the ones with the most obvious things wrong. And they responded quickest.
Around 250% growth means about three and a half times the starting point. The figure covers calls and direction requests attributed to the business profile, per location. It is not revenue and it is not a ranking position.
- The estate average was the problem, not a symptom of it. It was stable throughout the period when several locations were invisible.
- Comparability had to come before optimisation. Locations measured differently cannot be ranked against each other, however confident the table looks.
- Bulk correction turned a multi-week job into an afternoon. This is the difference between the work happening and being for good deferred.
- Roughly 250% growth in calls and direction requests per location, against the baseline recorded before the programme began.
What this does not claim
Per-location results varied much. The headline figure describes the outcome per location against its own baseline, not a uniform result across every site in the book.
Locations that were already performing well moved far less. Most of the aggregate movement came from sites that had specific, fixable faults.
The figure covers profile-attributed calls and direction requests. Client revenue is not measured here and is not claimed.
Questions about this account
Why is an estate average the wrong instrument?
It is stable by construction. Individual failures are diluted by everything else in the set. So a location can be invisible for months without the number moving.
What does a common frame mean?
The same grid geometry, spacing and keyword set applied to every location. So the ranking between them reflects performance rather than measurement differences.
Why group locations before ranking them?
A mid position in a dense city is often a better result than a top position in a thin market. Ranking them together rewards the easier market.
Did every location improve by the same amount?
No, and the headline should not be read that way. Most of the aggregate came from sites with specific, fixable faults.
How does this change client reporting?
The method gets stated in the document. This is what allows a client to check the figures rather than take the ranking on trust.
What is the earliest warning sign in a portfolio?
A location that has been quiet for several periods while the average holds steady. That combination is what this measurement is designed to show.
Start with a baseline of your own.
Every result above depends on a measurement taken before anything changed. $5.33 per location per month, all 42 tools.