The Short Version
TL;DR
→Star average above ~4.3 is table stakes. The four metrics that do the work in 2026: reviews per completed job, median review word count, 48-hour response rate, and platform distribution.
→Response rate has been a heavily-weighted LSA ranking factor since late 2024 — Google never announced it. Most teams still don’t track it because their dashboard doesn’t surface it.
→The three-tier rebuild: Easy — FSM automation + response templates. Medium — content prompting + platform split + response cadence. Hard — named review-ops owner + weekly dashboard + monthly reconciliation to jobs and revenue.
Response Rate Became an LSA Ranking Factor 18 Months Ago
In late 2024, Google quietly reweighted its LSA algorithm. Response rate joined the top tier of ranking factors alongside review count and star rating — no press release, no dashboard change, just a shift in how leads flow. Contractors who responded fast started getting more; those who didn’t got fewer and blamed “the market.”
Response within 5 minutes = an 8x conversion lift over a 30-minute wait.
Boomcycle Digital 2026 LSA analysis · Ferocious Media 2026 review-rank study
The gap that opened isn’t about who has the best reviews — it’s about who has the review machine. Companies with a real one track four metrics no reputation platform surfaces by default, put a named person’s scorecard on response rate, and read the last 20 reviews every month to feed ad and landing-page copy. They’re a small minority — and they’re the ones showing up in LSAs, ChatGPT, and the local pack right now.
Sources: Boomcycle — 2026 LSA Ranking Factors · Ferocious Media — Reviews First, Speed Second
From the team at DUO
The four-metric review dashboard is the single most common build we run for clients — and the first thing that gets forgotten when nobody owns it. If you want us to build it for you (or rebuild the FSM automation so it actually feeds the four metrics), we’ll set it up — and run one month of the dashboard free so you can see the shape of the data before deciding to own it internally.
Build my review machine →Why Most Review “Machines” Are Just Review Collection
Collecting reviews and running a machine are different systems. Here’s the gap, the diagnosis, and the three-tier build to close it before Q3 ends.
The problem
Most home services companies have a “reputation management platform” that fires a review request 24 hours after a job. That’s collection, not operations. Nothing measures response rate, tracks median word count, splits platforms, or reconciles review velocity against jobs completed. The dashboard the marketing director sees shows total reviews and star average — the two metrics that stopped mattering two years ago.
| Your dashboard probably shows | What actually ranks you now |
|---|---|
| Star average (4.8★) | Reviews per completed job (25%+) |
| Total review count (523) | Median review word count (50+) |
| Reviews collected this month | Response rate within 48 hrs (90%+) |
| Google reviews only | Platform mix (60/30/10 target) |
The diagnosis
You likely have a review collection system, not a machine, if any of these are true:
- ×You can’t say what your median review word count is.
- ×You don’t know your response rate within 48 hours.
- ×More than 80% of your reviews are on a single platform.
- ×You report “reviews collected” monthly but not “reviews per completed job.”
- ×No single named person owns the review dashboard.
A working machine has three layers — automation, process, ownership. Most companies run layer 1 at ~60% and layers 2–3 don’t exist. That gap is why the industry’s star average keeps rising while its ranking keeps dropping.
The fix — pick your level
The build is a stack: Level 1 has to work before Level 2 makes sense, and Level 2 before Level 3 is worth the headcount. Start where you are.
Level 1 — This week
Fix the FSM automation. Fix the response gap.
3 hours setup + 15 min/day. FSM admin + one CSR or ops person.
1Turn on post-job review requests in your FSM. Jobber, ServiceTitan, and Housecall Pro all support it natively — fire 24 hours after completion. If it’s already on, verify it fires on every job, not just ones tagged “request review.” The usual breakage is a missing tag on ~40% of jobs.
2Write three response templates. 5-star (specific + thank + name the service); 3–4 star (thank + solve + invite back); 1–2 star (empathize + take offline + resolve). Save them so anyone can respond in under 60 seconds.
3Assign one response owner. Not a task — a KPI on their scorecard: 90% of reviews responded to within 48 hours. This is the single biggest LSA move most companies never make.
Level 2 — This month
Build the content layer and split the platforms.
1 day setup + ongoing cadence. Marketing person or agency.
1Rewrite the request to prompt for content. Not “leave a review” — “Would you mention the specific service we did and how fast we got there?” Target 50+ median words. Density is what feeds AI engines the keywords that surface you.
2Split requests across platforms. Google 60% (feeds Gemini + AI Overviews), Yelp 30% (feeds ChatGPT), BBB/trade 10% — or ask on whichever platform the customer mentions first. Google-only = a thin signal across half the AI surface.
3Put response rate in the monthly report. The actual % within 48 hours, charted monthly. Below 85% for two months is a scorecard conversation; 90%+ is a moat that compounds.
Level 3 — This quarter
Build the review-ops function.
Named owner + weekly dashboard + monthly reconciliation. 1 week setup, 3 hrs/month.
1Assign a named review-ops owner. Not the marketing director (too loaded), not a CSR (wrong altitude) — a fractional CMO, an expanded CSR lead, or a $15–25/hr part-time hire whose primary KPI is the four metrics.
2Build a weekly one-page dashboard. Four rows: reviews per completed job (25%+), median word count (50+), 48-hour response rate (90%+), platform mix (60/30/10). Sent to owner + marketing + ops every Monday; two weeks below target = an explanation.
3Monthly reconciliation — reviews as market research. Read the 20 most recent reviews together (30 min). What services get mentioned? What language do customers use? Feed it into ad creative, landing pages, and service pages. Almost nobody in home services does this yet.
Do this next
- ☐Run the five diagnosis questions. Any “yes” = you have a collection system, not a machine.
- ☐This week: verify FSM review automation fires on every job, not most.
- ☐Write the three response templates. Assign one named owner with a 90%-within-48-hours scorecard.
- ☐Rewrite the request prompt to ask for content: “mention the service we did and how fast we got there.” Target 50+ words.
- ☐Split the platform mix: Google 60 / Yelp 30 / BBB or trade 10, rotated through FSM automation.
- ☐Calendar a monthly 30-min reconciliation — marketing + ops + review owner read the last 20 reviews together.
