LMH Agency
Home services contractors, Meta alongside search and messaging.
33.0 of 100 on disclosed practice. Read 12 September 2026, restated quarterly.
2 ads running under their own Page. Newest both 27 May.
A count for one date. How we checked.
| Criterion | Score | What the published evidence shows |
|---|---|---|
| Causal measurement | 12 / 25 | Names two third-party measurement tools, including call scoring, which is genuine lead-outcome measurement feeding quality back into the account. No reconciliation to verified revenue, no holdout. |
| Conversions API and event quality | 4 / 20 | Lead forms and retargeting imply the pixel. No Conversions API anywhere on the page, no offline events. |
| Objective coverage | 9 / 18 | The clearest funnel-stage reasoning in the set: awareness at the top, leads in the middle, conversions at the bottom, each with a stated role. Instant forms and retargeting named. Calls and Advantage+ absent. |
| Creative iteration | 3 / 15 | A bare mention of A/B testing with no structure and no number. Two ads found running, both from late May. |
| Spend band, published | 0 / 12 | No budget, minimum or fee published anywhere. |
| Claim hygiene | 5 / 10 | Neutral. The page publishes no performance figure at all, so there is nothing stated dishonestly. |
Source: lmh.agency, Meta ads for home services page. Band definitions on the method page.
What they are strongest on
They think in funnel stages and say so, which fits how Meta actually works better than most of the field.
Where they are weakest
Nothing published about tracking, cadence or cost.
Read from lmh.agency, Meta ads for home services page on 12 September 2026, against published criteria. Page changed since? The corrections page gets it re-read.
- 16 points sitting unclaimed on conversions api and event quality. Document server-side events, and whether booked jobs go back to Meta rather than only into your client’s CRM.
- 13 points sitting unclaimed on causal measurement. Publish the design of a holdout, geo test or lift study you have run, including what was randomised and over how long.