You know you're losing customers—they just stop coming back, quietly, without a complaint or a cancellation. But here's the question almost no local business owner can answer precisely: how long does someone have to be gone before they count as "lost"? Get that number wrong and your win-back campaigns either nag loyal regulars or chase people who left months too late. Get it right and you unlock the single cheapest source of revenue you have.
Why "Lapsed" Needs a Real Definition
Most owners run on gut feel. "I haven't seen Maria in a while" is not a system. Without a concrete definition of lapsed, you can't build a list, can't time an outreach, can't measure whether your win-back efforts worked, and can't tell the difference between a customer who's overdue and one who's gone for good.
A clear definition turns a vague worry—"are we losing people?"—into a managed pipeline you can act on. It's the foundation of any reactivation program. If you don't know who's lapsed, you have no one to win back. Our customer reactivation guide covers the full playbook; this post is about getting the first, load-bearing definition right.
The cost of getting it wrong
Define "lapsed" too aggressively—say, 30 days for a business where people naturally buy every 90—and you'll bombard active customers with "we miss you!" messages that feel needy and out of touch. Define it too loosely—wait a year for a coffee shop—and the customer has already built a new habit somewhere else, far past the point where a text could pull them back. The right threshold is specific to your business, and finding it is mostly arithmetic.
The Core Concept: Purchase Frequency
The key to defining lapse is your average purchase cycle—the typical gap between one purchase and the next for a returning customer. A customer becomes "at risk" when they've gone meaningfully past that cycle, and "lapsed" when they've gone well past it.
Think about how differently this plays out across business types:
- A coffee shop regular might come in three times a week. Two weeks of silence is already a red flag.
- A hair salon client typically returns every 6–8 weeks. Twelve weeks out is a warning sign.
- An HVAC company might see a customer for a tune-up once or twice a year. "Lapsed" could mean 18 months.
- A roofer may serve a household once a decade. The frame is completely different—reactivation there is about referrals and adjacent services, not repeat purchase.
The lesson: there is no universal number. Anyone who tells you "90 days" is a generic guess. You have to derive yours.
How to Calculate Your Own Lapse Threshold
You don't need a data scientist. You need your transaction records and about an hour.
Step 1: Find your average repurchase interval
Pull a sample of customers who've purchased at least twice. For each, calculate the number of days between consecutive visits. Average those gaps. That number is your typical purchase cycle.
If you have point-of-sale or booking software, export the last 12–24 months of transactions with customer identifiers and dates. Even a rough sample of 50–100 repeat customers will give you a defensible average.
Step 2: Add a buffer for normal variation
Customers don't run on a perfect schedule. A salon client who's "every six weeks" will sometimes come at five and sometimes at nine. So your lapse threshold isn't the average—it's the average plus a cushion that accounts for normal lateness.
A practical rule: multiply your average cycle by roughly 1.5 to 2 to define the at-risk window, and by 2 to 3 to define fully lapsed. A salon averaging six weeks would flag customers as at-risk around 10–12 weeks and lapsed around 14–18 weeks.
Step 3: Look at the distribution, not just the average
Averages can mislead if you have two very different groups—say, monthly regulars and once-a-year visitors. If that's you, segment first. Define lapse separately for each group, because lumping them together produces a threshold that's wrong for both. This is the logic behind RFM analysis, which sorts customers by recency, frequency, and monetary value so you can treat the groups differently.
At-Risk vs. Lapsed vs. Lost: A Three-Stage Model
A single "lapsed" flag is better than nothing, but the businesses that win back the most revenue use a staged model. Each stage calls for a different message and urgency.
At-risk. The customer is past their normal cycle but not by much. This is your highest-value moment to act, because the relationship is still warm and the habit isn't broken. A gentle, helpful nudge here often prevents the lapse entirely. Think: a reminder, a "time for your next service?" prompt, a small convenience offer.
Lapsed. The customer is well past their cycle—roughly two to three times the normal interval. The relationship has cooled and they may have drifted to a competitor or simply forgotten. This is the classic win-back zone, where a stronger reason to return—often an offer—does the heavy lifting.
Lost / dormant. The customer is so far past their cycle that a return is unlikely from a single campaign. They're not worthless—a great offer or a major change at your business can still revive some—but your expected response rate is low, so you spend less effort and lead with your strongest incentive.
Why staging beats a single flag
Staging lets you match effort to probability. You invest the most thoughtful, lowest-cost outreach in at-risk customers (where a small touch has big payoff), reserve your real offers for the lapsed (where you need a reason), and run occasional, low-effort blasts at the dormant. It's the difference between a scattershot "we miss you" email and a system. For the related question of where to draw these lines for repeat-purchase loyalty programs, see loyalty vs. reactivation.
Common Mistakes in Defining Lapse
Using a competitor's or vendor's default number. A 90-day default baked into some marketing tool has nothing to do with your purchase cycle. Always derive your own.
Ignoring seasonality. A pool service or a landscaper has a natural off-season. A customer who's "gone" for the winter isn't lapsed—they're seasonal. Build your definition around the in-season cycle, and time win-back for the start of the next season. Seasonal win-back campaigns need their own calendar.
Treating one-time buyers as lapsed regulars. Someone who bought once and never returned isn't following a "cycle" at all—they may have been a one-off (a tourist, a one-time emergency, a gift purchase). Win-back logic for them differs from a true repeat customer who's slipping. Segment them out.
Forgetting to update the number. Your purchase cycle shifts as your business changes—new services, new pricing, new clientele. Recalculate your lapse threshold once or twice a year so it stays honest.
Turning the Definition Into Action
Once you have a number, the workflow writes itself:
- Tag customers automatically as they cross your at-risk and lapsed thresholds, based on their last transaction date.
- Trigger the right message for each stage—a soft reminder at at-risk, an offer at lapsed.
- Send by the channel they prefer—email, SMS, or both—and measure which brings them back. See email vs. SMS review requests for the channel tradeoffs that apply here too.
- Track reactivation rate by stage so you learn where your effort pays off.
This is exactly the kind of date-based segmentation Revive Local automates—it watches your last-visit dates, flags customers as they cross the thresholds you set, and fires the right win-back message at the right moment, so you're not manually scanning a spreadsheet wondering who hasn't been in lately.
The definition is the unlock. The moment you can name who's lapsed, you've turned a quiet, invisible leak into a list you can act on—and lapsed customers are far cheaper to bring back than new ones are to find.