A 4.8 on Google and a 4.3 on Facebook: rebuilding the reputation layer a trade business is actually judged on

The Google profile was in good shape at 4.8. The Facebook page was not, sitting materially lower at around 4.3, and Facebook was where a meaningful share of the local audience was being reached. So the strongest rating was on the surface a buyer might never see, and the weakest was on the surface the business was paying to put in front of people. One negative review was being contested through Google on a policy basis with no guarantee of an outcome. Bing Places was queued to sync from the Google profile and waiting on approval. Yelp had not been touched at all. Nothing was systematically asking satisfied customers for a review on either platform, so the ratings were being set by whoever felt strongly enough to write unprompted.

The paid social was sending local buyers to a Facebook page rated around 4.3, while the 4.8 on Google sat on a profile those buyers had no particular reason to open.

1. 01

   A review is disputed only where there is a clear factual or policy basis

   The test is narrow on purpose. A review that misstates a verifiable fact, or that breaches the platform's own content policy, is worth contesting. A review that is simply unfair is not. And a dispute never involves exposing private personal details about the reviewer or making an accusation about them, because a business that does that loses the argument in front of every future reader even when the platform sides with it.

2. 02

   The weakest rating is the one your media points at

   Reputation is not an average across platforms. It is whatever the buyer happens to be looking at when they decide. A 4.8 does not compensate for a 4.3 if the campaigns land on the 4.3.

3. 03

   A dispute is slow and uncertain; review velocity is neither

   Contesting a review takes weeks and may fail. Asking twenty satisfied customers is under the business's own control and moves the average in a way no single removal does. The two are not alternatives, but only one of them is a plan.

4. 04

   Asking for a review requires proof that the job happened

   The reason crews do not ask is that the ask feels like an imposition with nothing attached. A job-site photograph supplied that week turns the request into a follow-up on something the customer can see, which is why the photo supply line and the review motion are the same system.

5. 05

   Directory coverage is not equal-cost, so it is not equal-priority

   Bing Places syncs from the Google profile at close to zero effort and was simply pending approval. Yelp requires real work to set up and real work to maintain, and had not been started. Treating them as one line item on a checklist hides a large difference in what each one costs to do properly.

We baselined rating, review count and review recency per platform rather than in aggregate, then wrote a response standard: who responds, within what window, and in what shape. The dispute rule was written down as a test rather than left to judgement in the moment, and one review was contested through Google on a policy basis with the outcome treated as uncertain from the start. A review request motion went to both Google and Facebook at job completion, deliberately on both rather than only on the stronger platform. Underneath it runs a daily call with the client's marketing contact that produces one postable asset within two hours, targeting five posts a week with three counted as an acceptable week, fed by at least one job-site photograph supplied by the client each week. Bing Places was synced from the Google profile and left pending approval; Yelp was queued and honestly recorded as not started.

1. 1

   Weeks 1–2

   Reputation baseline

   Rating, review count and recency captured separately for Google, Facebook, Bing and Yelp

   Owner: OmniFlow

2. 2

   Weeks 2–3

   Response standard and dispute rule

   Written response shapes, owner, response window, and the narrow test for contesting a review

   Owner: OmniFlow + client

3. 3

   Week 3

   Contested review filed

   One review disputed through Google on a policy basis, outcome treated as not guaranteed

   Owner: OmniFlow

4. 4

   Week 3 onward

   Review request motion

   Request at job completion on both Google and Facebook, not on the stronger platform only

   Owner: OmniFlow + client crews

5. 5

   Ongoing

   Daily asset cadence

   One postable asset within two hours, five per week targeted, three an acceptable week

   Owner: Client marketing + OmniFlow

6. 6

   Weeks 4–6

   Directory coverage

   Bing Places synced from the Google profile, pending approval; Yelp queued and recorded as not started

   Owner: OmniFlow

7. 7

   Weeks 5–8

   Speed-to-lead instrumentation

   Response clock on every inbound, median time to first response reported alongside rating

   Owner: OmniFlow

Report this engagement on the two ratings side by side and on review recency, never on a blended score, because a blended score is exactly the view that hid the problem. Pull the Google figures from Google Business Profile Insights and the Facebook rating from the page itself rather than from any dashboard that reconciles them. The outcome of the contested review is reported as filed or removed, with no claim made before the platform decides. Median time to first response is the third number and belongs next to the other two, because in this trade the rating gets the call and the response time keeps it.

Measured in Google Business Profile Insights, the Facebook page rating and JobTread, window to be confirmed at publication. Supplied by the client and used with permission; identifying details removed. Cited: Invoca *Home Services Lead Conversion Benchmarks 2026*; BrightLocal *Local Consumer Review Survey*; LocaliQ *2025 Home Services Benchmarks*. ---.

These are the surfaces this engagement is run and measured from, shown with representative figures built around the benchmarks cited on this page. Every account we run reports into views like these, and you keep ownership of all of them.

2 of the figures on this engagement are not published yet. Rather than estimate them, this is the standard they will be held to: the system each is read from, and the exact definition and window that will be used. Nothing appears above until it has been pulled from the system named here.

We filed the contested review before the request motion was producing any volume. That is the wrong order. A dispute is slow, uncertain and outside our control, and the weeks spent waiting on it were weeks in which new legitimate reviews would have moved the average further than a removal ever could. On the next engagement of this shape the request motion goes live first and any dispute is filed underneath it, not instead of it. We were also slower than we should have been to treat Facebook as a first-class surface rather than as a secondary one.

Real client work. Every figure is measured in a named system and carries the window it was measured over. See all our work

Full citations for everything cited on this page, with the sample and period each study covers, so you can go and read the original.