The First 100 Reviews Playbook: From Zero to Untouchable

A homeowner searches for a plumber, an HVAC tech, or a roofer. Two shops come up. One has 214 reviews and a 4.8 average. Yours has three. She does not call to ask which shop does better work. She calls the 214. That gap closes with a system, and this post is that system: the foundation covered in the 5-Star Review Engine pillar guide, sequenced here for a shop starting near zero and aiming at its first 100.

Three moving parts: ask on every job the same way, build velocity in steady ramps instead of bursts, and respond to every review on a set cadence. Get those running and 100 becomes a date you can estimate from your job count.

Why zero reviews costs you jobs you never hear about

At zero to ten reviews, you are not losing arguments. You are being filtered out before the argument starts. A homeowner who has never heard of you has one shortcut for judging you: the star rating and review count next to your name. No count means she picks the shop that has one.

None of that requires being the biggest shop in town — just a rating at or above four stars, an established-looking count, and visible responses. Concrete takeaway: your review count is a production number, like jobs completed. It responds to a process, not luck.

The job-by-job ask system: every job, same ask, same timing

The entire engine is one standard operating procedure (SOP) — a written sequence your whole team runs the same way, every time — with no judgment calls inside it. The tech does not decide who gets asked. Every completed job triggers the same touches:

• Touch 1 — the in-person ask at completion: as the tech walks the customer through the finished work, the tech asks out loud, in the same words every time, and says a text with the direct link is on its way. Asked in person first, that text is expected, not spammy.
• Touch 2 — the text with the link, within the hour: one short text with your direct Google review link, sent while the job is still fresh. No long explanation, no second link, no survey first.
• One reminder, then stop: a single reminder three to four days later, worded exactly like the first ask. After that, the job closes. Chasing further trains customers to ignore your number.

Record every ask in your CRM (customer relationship management software — the system where you store customer records and job history): ask sent, reminder sent, review received. What lives in the tech's memory disappears by Friday. Concrete takeaway: the ask is part of completing the job, like the invoice and the cleanup. A job is not closed until the ask has gone out.

The velocity ramp: 0 to 10, 10 to 40, 40 to 100

Review velocity is new reviews per week. Four a week, every week, looks alive. Thirty in one week and none for two months looks like a campaign. Work the ramp in three phases:

• Reviews 0–10 — prove the ask works: run the two-touch system on every job for two weeks and fix whatever breaks. Tech forgetting? Put the ask on the job-completion checklist. Link not opening? Test it on your own phone. Most shops find their ask rate, not their customers, was the problem.
• Reviews 10–40 — add the backlog, carefully: alongside new jobs, work your past-customer list in weekly batches of twenty to thirty, same ask and link. A steady stream from real past jobs is legitimate; a sudden flood is not.
• Reviews 40–100 — hold the cadence: by now new jobs alone should carry you. Keep backlog batches going only until new-job asks sustain your weekly target, then retire the backlog and let the engine run on current work.

Example: a two-truck HVAC shop completes 24 jobs a week. If one customer in four who gets asked leaves a review, that is roughly six a week — the first 100 arrive in about four months. Plug in your own job count and rate; your numbers set your date. Concrete takeaway: set a weekly target from job volume, check it every Monday, and treat a missed week as a process break to find.

The response cadence: every review, answered on schedule

Asking earns the review. Responding is what future readers judge — 89% of consumers expect owner responses (BrightLocal 2026), and your replies read as proof a real owner stands behind the work. Run responses as a cadence:

• Twice a week, same days: pick two fixed slots — say Tuesday and Friday mornings — and clear every unanswered review. No review sits longer than three to four days.
• Thank specifically on positive reviews: name the job or the tech, in one or two sentences. A response that could sit under any business's review reads as automation, because it usually is.
• Slow down on critical reviews: acknowledge the concern, say what you will do to look into it, and move the detail offline to a named phone number. Never argue in public or blame the customer. You are writing for the next homeowner, not the reviewer.

Concrete takeaway: assign response duty to one named person with a backup, on the same calendar as payroll. Reviews answered by "whoever gets to it" get answered by nobody.

What never goes in the system

• Never pay for reviews or offer anything for them: no discounts, gift cards, prize entries, or free add-ons tied to a review. That violates Google policy, and reviews earned that way can be removed.
• Never screen customers before the ask: sending the link only to customers who seem pleased, or routing people based on how the job went, is review gating. Asking every customer the same way is both the policy and the system. Some reviews will be four stars — a profile with honest fours and owner responses beats a suspicious wall of fives.
• Never write, buy, or borrow reviews: no staff reviews, no family reviews presented as customers, no swaps with the shop across town. One exposed fake can cost the whole profile.
• Never change the ask based on the customer: once techs pick who "seems like a reviewer," your ask rate collapses and your data goes blind. Same ask, every job, is what makes the numbers predictable.

Compliance notes: texting the asks

• Registration first: A2P (application-to-person messaging) with 10DLC (10-digit long code) registration is required before sending business texts — approval typically takes 1–7 days and costs roughly $15–$20 one-time.
• Consent and opt-out: text customers who agreed to be contacted, and include opt-out language in the first text — the ask script above carries it.
• Quiet hours: respect TCPA (Telephone Consumer Protection Act) quiet hours — no marketing texts before 8am or after 9pm in the recipient's local time. An 8:40pm job close gets its text the next morning.
• Reviews: ask every customer the same way, offer nothing in exchange, and never filter who receives the ask. This section is general information, not legal advice.

Related guides in this series

This playbook sequences the 5-Star Review Engine pillar guide into a first-100 plan for shops starting near zero. When the manual cadence starts eating office time, How to Automate Google Review Requests After Every Job shows how to trigger the same touches automatically. And How to Ask for Google Reviews: Timing, Wording, and Scripts goes deeper on ask wording.

The done-for-you version

This post is System 4 from the Leveraged Owner Starter Pack — the done-for-you version with the full SOP, every script, the worksheets, and screen-by-screen setup instructions for all 8 systems. It includes this first-100 engine pre-written — ask scripts, reminder sequence, backlog tracker, and response routine — ready for your business name and review link.

Stats sourced as labeled: review behavior figures from BrightLocal 2026 (68% won't use a business under 4 stars; 97% read reviews; 71% use Google to find local businesses; 89% expect owner responses). Worked examples labeled as examples; not real customer results.

Keep building your systems

Related guides from the Leveraged Owner blog: