Automated NPS surveys from your RealGreen visit data

Can you run automated NPS surveys off your RealGreen visit data? Yes, and it is one of the simplest automations we build. The visit record is the trigger: a customer gets serviced, a one-question survey goes out on a schedule you control, detractors route to your office before they route to Google, and promoters get pointed at your review link. You do not need standalone survey software you will abandon by March. Set it up once and it runs all season, and what you get back is a retention early-warning system that costs almost nothing to operate and tells you things your cancel report only admits after the money is gone.

Why bother with NPS when you already ask for reviews?

Because reviews and NPS answer different questions. A review request is public marketing: it hears mostly from people willing to say something nice with their name on it, and its job is to build your Google profile. We covered that machinery in our guide to automating Google reviews. NPS is private diagnostics. It hears from the quiet middle of your base: the customer who is not angry enough to call but is one brown patch away from shopping your competitor's postcard. Reviews tell prospects you are good. NPS tells you which customers are about to leave, which techs are quietly excellent, and which branch has a problem nobody has reported yet.

They also feed each other. The promoters your NPS survey identifies are exactly the people your review automation should ask next, and the detractors are exactly the people it should never ask.

What should the survey actually ask?

The classic form works: "How likely are you to recommend us to a friend or neighbor, 0 to 10?" Then one open follow-up: "What's the main reason for your score?" That is the whole survey. Every additional question costs you response rate, and response rate is the entire game. A twelve-question survey with a 4% completion rate teaches you less than a one-tap score with a 25% response rate, because the twelve-question version only hears from people with strong feelings and free afternoons.

Text is the right channel for most residential bases: the customer taps a number and is done in ten seconds. Email works as the fallback for customers without a mobile number on file.

When should the survey go out?

Not after every visit. Surveying five or six times a year annoys people and trains them to ignore you. Tie the sends to moments in the season when the answer is actually useful:

• After round two. First impressions have formed, early service problems have surfaced, and there is still a whole season left to fix them.
• Mid-season. Usually after the summer round, when heat stress and weed pressure make unhappy customers unhappy out loud.
• Before renewal letters drop. A score captured in early fall tells you who to call before you ask them for next year's money.

Then add suppression rules, because timing is where homegrown programs embarrass themselves: no survey within two weeks of an open complaint, no survey to a past-due account (asking "how are we doing" while dunning someone reads badly), and never more than one survey per 60 days no matter what triggers fire.

How does the automation actually work?

Trigger: a visit posts as complete in RealGreen for the rounds you picked, and the customer clears the suppression rules. Action: the survey text goes out the next morning, the score writes back to the customer record, and the routing rules below take over. Math: 1,800 active customers × 70% reachable by text × 25% response ≈ 315 scores per wave, roughly 900 a season. All illustrative: plug in your own base and response rates. That sample is big enough to read by branch and by tech, which is where the value lives.

Because visit data comes over on a nightly sync, next-morning delivery is the practical standard here, and it is fine: nothing about an NPS survey needs to arrive within minutes of the truck leaving. The survey tool itself is not the constraint either: we can connect RealGreen to pretty much any platform that has an API, a Zapier connection, or a native integration, so run the surveys wherever your office already works.

What do you do with the answers?

Detractors (0-6)

This is the reason the program exists. A detractor score should create an office task the same day, with the score and the customer's comment attached, and a human should call within one business day. That call is a save opportunity you would otherwise never get: most detractors do not complain, they just fail to renew. The full complaint-to-resolution loop is its own machine, and we wrote it up in closing the loop on service complaints.

Passives (7-8)

Leave them alone. Do not send passives a review ask, and do not send them a coupon as an apology for a seven. A seven is a customer who is fine.

Promoters (9-10)

Route them to the review ask, and after that, the referral ask. A promoter who just told you privately that they would recommend you is the best possible candidate to do it publicly, and the sequencing matters: score first, review ask second converts noticeably better than a cold review request.

What do branch-level and tech-level scores tell you?

A company-wide NPS number is trivia. The management value shows up when you segment. By branch, a persistent ten-point gap usually means a production or supervision problem, not a customer problem. By tech, the pattern is more interesting: over a season you will find one or two techs whose routes score consistently high. Those are the techs customers already trust, which makes them the natural candidates for a technician lead program. Put the segmented score on the same one-page report as your weekly numbers so it gets looked at, not filed.

One honest caveat on tech-level scores: sample sizes are small. A tech with eleven responses and one angry outlier is not a coaching case. Read tech scores over a full season, not a month, and treat them as a conversation starter rather than a scoreboard.

Where do NPS programs go wrong?

Three failure modes account for most dead NPS programs. Nobody owns detractor follow-up, so scores pile up unanswered and the survey becomes a way to document churn instead of prevent it. The score becomes wallpaper: it goes on a dashboard, nobody's Monday changes, and six months later the program is quietly off. Or the survey grows: someone adds questions about the website and the invoice format, response rates halve, and the signal drowns. If you cannot commit to a human calling every detractor, skip NPS entirely and spend the effort elsewhere. An unanswered survey is worse than no survey, because the customer told you and watched you do nothing. Purpose-built automation platforms handle the mechanics: the harder part is the callback discipline, and that is on you.

The takeaway: your visit data already knows exactly when to ask each customer how it's going. One question, sent at the right three moments, with detractors called within a day and promoters routed to reviews, is a retention system disguised as a survey.

Start this month

1. Pick the three survey moments for your market: post-round-two, mid-season, pre-renewal.
2. Write the two-message survey (score question, reason follow-up) and set the suppression rules: complaints, past-due, 60-day cap.
3. Wire the trigger to completed visits from your RealGreen sync, with scores writing back to the customer record.
4. Assign detractor callbacks to a named person with a one-business-day standard.
5. Add branch-level and tech-level NPS to your weekly report and review it for a full season before making personnel judgments.