Review Request Email vs Text: The Honest Split Test

The owner swears by text. "Nobody opens email anymore," he says, waving his phone. The office manager swears by email. "Texting customers feels pushy, and email is free," she says, pointing at the invoice software that already sends receipts. They have had this argument four times this year. Neither of them has ever measured anything.

Why your opinion on this is worthless

Ask ten contractors whether email or text gets more Google reviews and you will get ten confident answers and zero data. The email camp points at cost: email is basically free. The text camp points at attention: texts get read in minutes. Both arguments sound right. Both are guesses about your customers, and your customers are the only ones whose behavior counts.

A residential HVAC customer who found you on Google behaves nothing like a property manager who sends you forty work orders a year. A 28-year-old homeowner texts like breathing. A 70-year-old landlord might never have texted a business in her life. The channel that wins for your shop depends on your list, your trades, and your towns. There is exactly one way to find out: run the test on your own customers and read your own numbers.

This post is part of the Pillar C "5-Star Review Engine" system, the part that turns finished jobs into fresh reviews. 97% of consumers read reviews and 68% won't use a business under 4 stars (BrightLocal 2026), so the channel that produces the most reviews per hundred jobs is worth knowing precisely. Here is how to know it in two weeks.

The two-week split test, step by step

You need about a hundred completed jobs to get a readable result. For most shops that is two to four weeks of work. Here is the setup:

• Split every completed job into two groups. Alternate them, one to email, one to text, in the order jobs close. Alternating beats hand-picking, because hand-picking smuggles your biases in. Odd job numbers get email, even numbers get text, done.
• Send at the same trigger point. Both groups get the ask at the same moment in the job lifecycle, for example the day the job closes and the invoice goes out. Timing is a huge variable, so lock it down. Our timing guide is Review Request Timing: Exactly When to Ask After the Job.
• Use the same review link in both. Your Google review short link, identical in the email and the text. Different links mean you cannot compare cleanly.
• Keep the ask equivalent. Same tone, same length of message, same sender name. You are testing the channel, not your copywriting.
• One ask per job, then stop. One follow-up nudge max, same for both groups. More than that and you are testing persistence, not channel.
• Log everything on one sheet. Job number, group (email or text), date sent, review posted (yes/no), date posted, star rating, word count of the review. A spreadsheet is fine.

That is the whole apparatus. No software to buy, no consultant needed. The discipline is in the alternating and the logging, not in anything fancy.

What to measure beyond open rates

Open rate is the vanity metric here. An opened email that produces no review lost to an unopened text that produced one. Measure what ends in reviews:

• Review conversion rate. Reviews posted per 100 asks, per channel. This is the headline number. Everything else is commentary.
• Time to review. How many hours from ask to posted review. Faster reviews mean fresher detail and better star ratings.
• Review quality. Average star rating and average word count per channel. A channel that produces three-sentence reviews beats one that produces "great service" every time, because detailed reviews sell the next job.
• Opt-outs and complaints. "Stop" replies on text, unsubscribes and spam complaints on email. A channel that wins on conversion but burns goodwill is not winning.
• Cost per review. Email is near free. Texts cost a few cents each plus your texting platform. Divide each channel's total cost by reviews produced. This is the number that settles the owner's "email is free" argument honestly.

Run the numbers at the end of the two weeks. You will have an answer with your customers, your jobs, your towns. That answer beats every guru's opinion, including mine.

The text script

Short, human, one link. Send it from the tech or the business name the customer recognizes, not a random short code. And the compliance block, because texts are regulated: A2P 10DLC registration (the carriers' registration system for business texting) is required before you send business texts, a process that typically takes 1–7 days with a roughly $15–$20 one-time fee. Include opt-out language in the first text you send a customer, and observe TCPA quiet hours: no marketing texts before 8am or after 9pm in the recipient's local time. This is general information, not legal advice. Our full texting templates live in Google Review Request Text Templates: 5 Messages That Get Reviews.

Why it works: it names the tech they actually met, it names the time cost honestly (60 seconds), and it doesn't beg. "Either way, thanks for the business" is doing quiet work there. It makes the ask feel optional, which is exactly why people do it.

The email script

Email gets more room, so use it: a real subject line, one line of job-specific detail, the link as a button-style line they can tap on their phone. Most people read email on their phones now, so write it like a text with a subject line.

The job-specific detail is the whole trick. "The water heater replacement" proves a human wrote this about their house, not a robot blasting a list. If your invoice software can merge the job type into the email, automate that field first. It is the highest-leverage personalization in the message.

When each channel wins

After dozens of shops run this test, patterns emerge. Your numbers are still the boss, but here is what to expect so you can read your results intelligently:

• Text usually wins on speed and residential. Homeowners answer texts fast, and the review gets posted while the job is fresh. Simple, one-visit jobs convert best here.
• Email usually wins on commercial and older customers. Property managers, facility contacts, and landlords live in email. So do customers who want the link on their desktop rather than their phone.
• Email wins when you lack texting consent. If you don't have a clear opt-in for texts, email is the compliant channel. Don't text people you have no business texting.
• Text wins on simple jobs, email on complex ones. A drain cleaning gets a text. A $14,000 system replacement with a story to tell gets an email, because the customer wants room to write the story.

The most common outcome of the split test is not a winner. It is a segmentation: text the residential one-visit jobs, email the commercial accounts and the big-ticket jobs. The test doesn't pick a channel. It draws you the map.

The rules that apply to both channels

Two rules govern every review ask you send, email or text, and they are not optional:

• Never filter unhappy customers before the ask. Asking only the smiling customers and quietly skipping the frustrated ones is review gating, and it violates Google's policy. Ask every completed job, both groups, no exceptions.
• Never pay for or incentivize reviews in ways that violate Google policy. No discounts for reviews, no gift cards for five stars, no "leave us a review and we'll knock $20 off." The ask is the incentive. Keep it clean.

These rules also protect your data. If you gate the ask, your split test is measuring your filtering, not your channels. Clean asks, clean numbers.

Example: reading a result sheet

Example: a three-truck electrical shop runs the test on 100 completed jobs, 50 email and 50 text, asked the day each job closed. Here is what an illustrative result sheet looks like. The numbers below are invented to show the math, not real customer results.

Email: 9 reviews from 50 asks, an 18% conversion rate, average 2.1 days to post, average 41 words per review, 1 unsubscribe, near-zero cost. Text: 14 reviews from 50 asks, a 28% conversion rate, average 5 hours to post, average 22 words per review, 2 STOP replies, about $4 in message costs.

Reading it honestly: text wins on volume and speed, email wins on review depth and cost. The shop's move is segmentation, text for residential service calls and email for panel upgrades and commercial accounts, then re-test quarterly because lists drift. Your sheet will look different. That is the point. Your customers, your answer.

Related guides in this series

The full review engine this test plugs into is the Pillar C "5-Star Review Engine" post. For the message wording itself, use Google Review Request Text Templates: 5 Messages That Get Reviews, and lock in your send moment with Review Request Timing: Exactly When to Ask After the Job.

The done-for-you version

This post is System 4 from the Leveraged Owner Starter Pack — the done-for-you version with the full step-by-step SOP, every script, the worksheets, and screen-by-screen setup instructions for all 8 systems.

Stats sourced as labeled: review figures (68% under-4-stars avoidance, 97% read reviews, 71% Google discovery, 89% expect owner responses) from BrightLocal 2026.

Keep building your systems

Related guides from the Leveraged Owner blog: