The Cleaning Company That Tracked Churn by Onboarding Source and Fixed a Leaky Funnel

A $1.8M residential cleaning company in the Mountain West had a churn problem that looked random. Some months they lost 2% of their recurring customers. Some months they lost 7%. The owner could not figure out why.

She had tried everything the cleaning business coaches recommend. A welcome folder. A 24-hour check-in. A 30-day follow-up. Her net promoter score was solid. Reviews were good.

When she finally broke churn data down by onboarding source instead of by month, the pattern was immediate. Customers who came from one specific lead source were churning at 3x the rate of every other source.

The cleaning business nobody tracks by source

Most cleaning operators track acquisition by source. Very few track retention by source. The assumption is that once a customer is in the book, the lead source does not matter anymore. Every customer goes through the same onboarding. Every customer gets the same service.

The data does not support that assumption. Google Ads cleaning clients churn differently than Facebook clients. Referred clients churn differently than both. The problem is the data is split across a CRM, a marketing platform, and a scheduling tool, so nobody connects the dots.

Neel Parekh of MaidThis put the stakes clearly: "50% of customers will switch cleaning companies after one bad experience. That number should scare you if you're in the cleaning biz." The leak is in the early experience, and the early experience varies by where the customer came from.

Text Sully: "Show me 90-day retention rates for recurring cleaning customers broken down by lead source, over the last 12 months."

The source that was burning through customers

When she sorted churn by source, Facebook lead-form clients were dying inside 90 days at a 41% rate. Google Ads clients churned at 18% in the same window. Referred clients churned at 9%. Website direct-form clients sat at 14%.

The shop-wide churn of 24% was the average hiding a very uneven distribution.

The 41% Facebook number was burning cash in two directions. Acquisition cost on Facebook was $62 per customer. Lifetime value on a 41% early-churn client was dropping below $800. She was spending $62 to acquire customers who were paying her $780 on average before quitting, a 12.6x ratio that sounds healthy until you subtract service cost.

Text Sully: "Calculate LTV to CAC ratio by lead source, using 12-month retained revenue."

What the onboarding calls revealed

She called the last 30 Facebook lead-form clients who had cancelled. Twenty-one answered. The pattern across those 21 calls was consistent.

The Facebook lead form had asked four questions. How big is your house. How often do you want it cleaned. What is your ZIP code. What is your name and number. The ad creative had promised a $99 first clean.

What the clients expected was $99 for a full deep clean. What they got was a $99 promotional rate on a standard clean, followed by a second visit at $189, then recurring at $165. Three out of four clients described feeling "bait and switched," even though the fine print was technically accurate.

This is the pattern every cleaning industry retention piece describes. Customerthermometer, Zenmaid, and Market Disruptors all name the same top churn driver: misaligned expectations set at the moment of booking. The source that sets bad expectations is the source that churns.

The Google Ads funnel that worked

Google Ads clients, who came in at 18% early churn vs. 41% on Facebook, had a different experience at the top of the funnel. They had searched for "recurring house cleaning" or a near-variation. They had landed on a page that listed actual pricing tiers, not a $99 promotional hook. They had filled out a longer form.

By the time they booked, they knew what the service cost and what it was going to include. The onboarding expectation-setting was already 60% done by the ad click.

This lines up with Google's own customer lifetime value guidance. Customers acquired through channels that match promise to delivery stay longer and spend more. The channel is not neutral. The channel is part of the onboarding.

Text Sully: "Show me new customers from each lead source in the last 90 days with their first-visit satisfaction score and their status at day 60."

The referral customers who stayed forever

Referred clients had a 9% early churn rate. On LTV, the referred cohort produced 16% higher 12-month revenue than the shop average, which tracks with Mention Me's published cleaning-specific referral data.

The mechanism was trust, not rewards. When a customer referred a neighbor, the neighbor came in already believing the service was good. The first clean did not have to sell them on the company. It had to not disappoint them.

That is a very different operational standard. The first-visit bar on a paid lead is "make a strong impression." The first-visit bar on a referral is "do not blow it." The second is much easier.

The owner stopped treating referrals as a nice-to-have and started treating them as her most important acquisition channel. She added a $40 referral credit on both sides. Referral volume in the next quarter doubled.

The complaint pattern she had been missing

Cross-referencing churn with complaint data surfaced the next issue. Customers who filed one complaint in their first 60 days churned at 63%. Customers with zero complaints in the first 60 days churned at 11%.

The pattern did not matter how the complaint was resolved. Once a customer had raised a concern in the first two months, the relationship was roughly half as likely to survive the year.

This mirrors what Customer Success Collective reports on early-tenure churn across recurring services: the first 30 to 90 days define the account. A single unresolved irritation, or a resolved irritation that left the customer feeling like they had to fight for it, sets a trajectory that is hard to reverse.

Text Sully: "Flag customers who filed a complaint in their first 60 days, sorted by whether they are still active."

The cleaner consistency problem

When she traced complaints further, most of them were about the cleaner changing. A customer had built rapport with a specific cleaner, that cleaner got rotated off the account, and the next visit felt different. Often the work was objectively fine. The customer's experience was not.

Zenmaid and Market Disruptors both name cleaner turnover and crew rotation as the single biggest driver of quality inconsistency complaints. The cleaner does not have to leave the company. They just have to leave that customer's home.

Her scheduling system had been optimizing for route efficiency, not for cleaner-customer continuity. She flipped the priority. The first rule of the schedule became "same cleaner, same house, same day-of-week if at all possible." Route efficiency dropped by 4%. First-year retention lifted by 11% on the cohort that got stable cleaner assignments.

Text Sully: "Show me customers whose primary cleaner has changed three or more times in the last 90 days."

The Facebook fix that changed the funnel

She did not shut off Facebook. She changed the ad.

The new ad promised "$149 first clean, recurring service from $139." The landing page listed pricing tiers and what each included. The form added a question: "Have you had a recurring cleaner before?"

Lead volume dropped 38%. Cost per lead went up 22%. But 90-day retention on Facebook-sourced clients moved from 59% to 84%. The channel started producing customers who were actually a fit for recurring service.

On total 12-month contribution, the new Facebook funnel produced more profit than the old one, with one-third the volume. The old funnel had been an expensive customer-churn machine dressed up as a cheap lead source.

What this means for your shop

If you are running a residential cleaning business, your churn number is probably hiding the same pattern. Some of your sources are profitable over a 12-month window and some of them are not, and the ones that are not are almost always the ones you have been praising for their "low cost per lead."

Cost per lead is the wrong number. Cost per 12-month-retained client is the right number. You can only calculate it when you can connect acquisition source to retention data, which almost nobody does.

Sully connects Jobber or Workiz or Housecall Pro data to your marketing sources and puts the math in chat. The prompts in this piece are actual queries you can run today. The answer tells you whether you should keep a channel, fix it, or turn it off.

For more on where AI fits into cleaning operations, see our AI for cleaning businesses guide, the customer reactivation playbook, the AI chatbot booking and qualification guide, and the AI review generation playbook.

Sources

• Zenmaid: How to Reduce Churn in Your Cleaning Business
• Market Disruptors: Cleaning Company Clients Cancelling
• Customer Thermometer: Reduce Customer Churn in Cleaning Industry
• Mention Me: Customer Lifetime Value Referral Marketing
• Zenmaid: Maid Service Expert Neel Parekh
• Check-N-Click: Onboarding is Customer Education
• Cleaning Business Academy: Onboarding New Cleaning Clients