The Lost-Customer Survey: Learning From Churn

A customer used you twice a year for three years, then stopped. Nobody called to complain. They just started using someone else — and you'll never know why unless you ask. Most contractors never ask. The ones who do discover their churn has a pattern, and patterns can be fixed. This post is part of the calendar and reactivation pillar: keeping customers is cheaper than finding them, and the survey is the diagnostic that tells you which part of the experience is pushing them out.

Who counts as "lost" (define it before you survey)

Not every quiet customer is a lost customer. Survey the wrong group and you'll drown in noise. Use these definitions:

• Recurring-service customers: canceled a maintenance plan, or missed two scheduled visits without rebooking.
• Repeat-work customers: used you 2+ times, then went silent for 18+ months (no service call, no tune-up, no quote request).
• One-time customers who never returned: 12+ months since the job, no review left, no second call — the "one and done" segment.

Exclude: customers with an open complaint (they're already telling you what's wrong — handle that directly), customers you fired, and anyone who opted out of your texts. Pull the list from your CRM monthly — it's a ten-minute export once the filters are saved.

The three questions (and why only three)

Long surveys get ignored. Three questions get answered — especially by text. Here is the survey, exactly as it should be delivered:

Each question does a different job. Question 1 (why you left) gets you the category: price, scheduling, communication, a bad experience, moved away. Question 2 (what could we have done better) gets you the fix in their words — often more specific and more useful than question 1. Question 3 (what should we fix) reframes them from a complainer into a consultant, which is why even unhappy customers answer it. "I read every one" is a promise you must keep — see the monthly ritual below.

Send it 30 days after the trigger event (cancellation, second missed visit, or the 18-month silence mark). Too soon and emotions are raw; too late and they've forgotten the details. One message only — this is research, not a campaign. If they don't answer, close the file.

The channel: text first, email as backup

Send the survey by text. Response rates on short text surveys run multiples of email for this audience — the customer already knows your number, the ask is personal, and answering takes thirty seconds. The compliance bar is the same as any business texting:

• Registered 10DLC business-texting registration number. A2P (application-to-person) 10DLC registration is required before sending business texts: roughly 1–7 days, about $15–$20 one-time.
• Opt-out language included in the first message, and instant suppression of every STOP.
• TCPA (federal telemarketing law) quiet hours: 8am–9pm recipient's local time. A survey sent at 10pm reads as spam no matter how polite the wording.

If a number is text-ineligible (landline, no mobile on file), send the same three questions by email with the subject line "Quick question from [Your Name] at [Business Name]." Expect a fraction of the response rate — but email costs nothing, so it's worth the send. This is general information, not legal advice.

Important distinction: this survey asks about their experience with you. It is internal feedback, not a public review request — so it is not review gating. Never use survey answers to decide who gets a Google review ask; the review ask goes to every completed job, full stop.

The monthly pattern ritual (this is where the value is)

Answers in a spreadsheet are trivia. Answers reviewed on a schedule are a diagnostic system. Once a month, spend 30 minutes on this ritual:

• Categorize every answer. Tag each response with one primary category: price, scheduling/wait time, communication, technician experience, quality of work, moved/life change, used someone else — no reason given, or positive (they'll be back).
• Count the categories. Ten responses with six "scheduling" tags is not ten anecdotes — it's one diagnosis: your booking process is losing customers.
• Pick one fix per quarter. Not five. One. If scheduling wins the count, the fix might be the arrival-window system, the confirmation cadence, or Saturday slots. Fix it, then watch next quarter's survey tags to confirm the needle moved.
• Reply to every respondent personally. "Thanks for telling me — we're changing [specific thing] because of feedback like yours." This closes the loop and, regularly, wins the customer back. People return to businesses that listened.

Keep the log for a year. The most valuable thing it produces isn't any single answer — it's the trend line. "Communication" tags dropping from 40% of responses to 10% after you installed the dispatch-text system is proof your systems are working, measured in the currency that matters: customers who stay.

What contractors actually learn (the common patterns)

Example: run the survey for six months at a 2-truck plumbing shop and the tags typically cluster like this. Price shows up, but it's rarely the winner — "used someone else, no reason given" and "scheduling" usually are. Dig one layer deeper on the scheduling answers and the pattern is almost always the same: the customer called with an urgent problem, got a four-day window, and called the shop that said "tomorrow morning." That's not a price problem — it's a speed-to-lead problem wearing a scheduling costume.

The second-most-common surprise: technician-experience tags that name no complaint about the work itself — "the tech was fine, the work was fine, but nobody told me what was happening." That's a communication fix, not a training fix, and it costs a text message, not a new hire. The survey's job is to keep you from solving the wrong problem with money.

And the "moved / life change" tags? Those aren't churn at all — they're list hygiene. Tag them, remove them from reactivation campaigns, and your reactivation numbers instantly get more honest.

Setting it up in an afternoon

The system has four moving parts: the saved CRM view (the three lost-customer definitions as filters), the survey text saved as a template in your texting tool, the monthly calendar reminder for the export and review ritual, and a one-page setup guide (standard operating procedure) covering the categories, the reply rule, and the quarterly fix selection. Write the setup guide once; your office manager runs the export, you spend 30 minutes a month on the categorization and the personal replies.

The win-back side effect deserves its own mention: a meaningful minority of survey respondents rebook within a month of answering — not because you sold them, but because being asked felt like being valued. Don't chase this; let it happen. The survey's primary job is intelligence. The rebooked customers are the bonus.

Related guides in this series

The lost-customer survey is the diagnostic layer of the Full & Calm Calendar pillar — it tells you which scheduling, communication, or service leaks are costing you repeat business, so your reactivation campaigns stop refilling a leaking bucket. Once the survey identifies the pattern, the fix usually lives in one of the calendar systems: confirmation texts, dispatch updates, or the post-job check-in.

The done-for-you version

This post is System 7 from the Leveraged Owner Starter Pack — the done-for-you version with the full step-by-step setup guide, every script, the worksheets, and complete setup guides for all 8 systems. The survey template, the categorization tags, and the monthly ritual checklist are pre-written and ready to install.

Stats sourced as labeled: no statistics used in this post.

Keep building your systems

Related guides from the Leveraged Owner blog: