A 600-client lapsed list looks like an opportunity in the dashboard and a liability in the inbox. Most salons either ignore it entirely or fire a single 20-percent-off win-back blast at the whole thing twice a year. Both approaches waste the actual asset, which is a small fraction of that list. The lapsed regulars worth real outreach sit underneath several hundred names who will never come back regardless: duplicates, one-time visitors, wedding-day clients, traveling guests, and people who already replaced the salon and forgot to delete the contact.
This article gives operators a triage method to sort the lapsed list before any message goes out. Four cuts, executed in order, against the existing CRM data. What remains is the working segment, typically 10 to 25 percent of the original lapsed file. That segment feeds the win-back sequence covered in the companion article. Triage is not optional. It is the difference between a retention system and a mass-email program.
Why running a single sequence against the whole list wastes the list
The mass-blast win-back has three failure modes layered on top of each other.
The first is the noise floor. Several hundred names on the list have no meaningful recovery probability. Duplicates, one-time wedding visits, traveling guests who flew home, clients who moved cities, clients who passed away. Sending to all of them inflates the send volume, drags down the response rate, and skews any subsequent analysis of what "worked."
The second is the discount signal. A blanket 20-percent-off message goes out, and a small fraction of the list rebooks. The salon counts the rebookings as wins. What the math hides is that the message also reached every full-price regular currently on the active list whose contact accidentally landed in the lapsed bucket, plus every current client who got the same email through her partner's account or a forwarded inbox. The discount signal quietly trains everybody to wait longer before rebooking next time. The win-back rebookings net positive on the campaign. The discount creep nets negative on the year.
The third is the service-mismatch problem. A lapsed balayage client at 14 weeks is not lapsed; she is on cadence. A lapsed single-process client at 14 weeks is two cycles past her window. Mixing them in the same blast sends both clients an inappropriate message. The balayage client reads "we think you have been gone too long" and feels misread. The single-process client reads a generic message and ignores it.
Triage solves all three before any sequence runs. The four cuts below are sequential. Each one removes a different category of unworkable names. What survives is the segment that warrants the real message.
The first cut: service category and service-specific cadence
The first cut is service category, because "lapsed" is defined by cadence, and cadence is defined by service.
For single-process color, lapsed begins at twice the normal cadence. A six-week regular is lapsed at twelve weeks. A five-week warm-tone client is lapsed at ten weeks.
For balayage and lived-in lightening, lapsed begins at twice the repaint cycle. A twelve-week balayage client is lapsed at six months. A fourteen-week client is lapsed at seven months. The midpoint gloss is a soft signal (a client who skipped the midpoint gloss is at higher risk of a full lapse) but it is not a lapse on its own.
For gloss-only clients, lapsed begins at twice the gloss interval. A four-week ash gloss client is lapsed at eight weeks. A six-week warm gloss client is lapsed at twelve.
The first triage operation is a CRM filter that pulls last visit date, last service category, and computes whether the gap exceeds twice the relevant cadence. Without this filter, the list is mixing apples and oranges and no downstream message can land cleanly. For the cadence definitions that drive this cut, see rebook cadence by service type.
The second cut: cadence cycles since last visit
The second cut sorts the survivors of the first cut by how many cadence cycles have passed.
The math is rough but defensible. Inside one cadence cycle past lapsed (so 12 to 18 weeks for a 6-week regular), recovery probability is highest. The client has likely just slipped, missed a rebook, or had a life event that disrupted the rhythm. A short, named message from the stylist often pulls her back.
Inside two cadence cycles past lapsed (18 to 24 weeks), recovery probability drops but is still real. The client has now skipped enough that she has either established a new routine with a different colorist or accepted home maintenance as her default. A stylist-level message can still work, but it has to be specific and personal. Generic salon-branded outreach to this segment is wasted breath.
Past three full cadence cycles (over 24 weeks for a 6-week regular, over 36 weeks for a balayage client), recovery probability collapses. Most clients in this bucket have a new stylist relationship and a new routine. They are reachable, but not with a standard win-back. They belong in a lower-frequency reactivation pool with a different kind of message, typically named and personal, run sparingly (twice a year at most), and never inside the same sequence that targets the one-to-two-cycle segment.
The three-cycle wall is the most important number in the triage. The salon that ignores it ends up running the same message against clients with wildly different recovery odds, which is the same failure as running a single sequence against the whole list, just at a smaller scale.
The third cut: lifetime value within each segment
After service category and cadence cycles, the third cut is lifetime value within each segment.
Lifetime value here is straightforward total spend across all visits, including retail. A high-LTV lapsed single-process regular who has spent $4,800 over four years and is now 12 weeks past her cadence is a different problem from a $280 single-visit client who came in 14 weeks ago for a corrective. Both are technically lapsed. Only one is worth a tailored outreach.
The defensible operation is to sort each segment by LTV descending and apply outreach in that order. The highest-LTV lapses in the one-to-two-cycle bucket get the first stylist message, ideally same week. The middle-LTV segment gets the message a week later. The low-LTV segment gets a lower-priority touch or, depending on the salon's bandwidth, no touch at all in this round.
LTV alone is not sufficient (a high-LTV client past three cycles is still a low-probability recovery) and recency alone is not sufficient (a recent low-LTV lapse is not worth the same effort as a recent high-LTV lapse). Both together, in that order, produce the working priority list. For the chair-side leverage that LTV math creates, see the cost of every lapsed regular.
The fourth cut: service-history flags that remove clients entirely
The fourth cut removes clients from the list entirely. These are not low-priority outreach; they are no-outreach.
Five categories qualify.
Documented service issues that were not resolved. Notes in the client record about a complaint, an adjustment that did not satisfy, or a stylist conflict. These clients did not lapse passively; they left. A generic win-back message confirms the salon did not listen the first time. If the salon wants to reach these clients, the message is different (acknowledgment, alternate stylist option, no marketing offer) and the workflow is documented separately. See the win-back article for the service-issue protocol.
Explicit do-not-contact requests. Honor them. The client opted out for a reason, and continued outreach is a regulatory issue in some jurisdictions and a reputational issue everywhere.
Chronic no-shows already moved to deposit-only booking. The salon has already made the decision that the relationship has constraints. Mass-blasting these clients with a win-back contradicts the policy and confuses the relationship.
One-time-event clients. Wedding parties, photoshoots, traveling guests, special-occasion visits documented at booking. These clients were never on a cadence to begin with. Their "lapse" is the salon's CRM not knowing the visit was one-time. Mark them in the system and remove them from every lapse-triggered workflow.
Clients whose contact information is wrong. The contact has bounced, the phone disconnected, the email returned. These names inflate the list and degrade the data. Either correct the contact or remove the client.
The fourth cut is the cleanup operation. Salons that skip it end up running outreach sequences against names that should not be in the workflow at all, which generates bad signal, wasted send volume, and occasionally an awkward reply from a client who told the salon two years ago that she had moved.
What the working segment usually looks like
After all four cuts, the working segment is usually 10 to 25 percent of the original lapsed file.
A salon with 600 names on the raw lapsed list typically ends up with 60 to 150 names in the working segment. Of those, perhaps 20 to 40 are in the high-priority bucket (one to two cycles past lapsed, high LTV, no service-history flags). That bucket is where the stylist-level outreach lives. The remaining 40 to 110 names sit in the medium-priority bucket and feed a slower, salon-level sequence run a week later.
The size of the working segment is part of the discipline. A salon that ends triage with 400 names in the working segment did not triage; it removed the obvious bad names and left every borderline case. The right size feels uncomfortably small. That is the signal that the cuts ran correctly.
This is the segment the win-back sequence is built to serve. Without triage, the sequence runs against the wrong list and produces the failure modes the triage was meant to prevent. With triage, the sequence runs against a list small enough to send personal messages on.
Where the triaged segments go
After triage, the lapsed list splits into four destinations.
The high-priority segment (one to two cycles past lapsed, high LTV, no flags) goes into the active win-back workflow. Stylist-level outreach, personal message, complimentary clinical add-on rather than a percentage discount. Two touches over 10 to 14 days, then stop.
The medium-priority segment (one to two cycles past lapsed, mid LTV) goes into the salon-level outreach. A short message from the salon's professional name, one touch, no follow-up unless the client replies. The economics of chasing this segment harder do not justify the time.
The low-frequency reactivation pool (past three cadence cycles) gets a different cadence entirely. One or two touches per year, calendared to high-leverage moments (early fall for color clients returning from summer, late winter for spring tonal changes). The message is named, personal, no marketing offer, no urgency. The point is to stay reachable, not to push a rebook.
The no-outreach pool (everything caught by the fourth cut) gets nothing. Mark the client record, document the reason, and leave them off every lapse-triggered workflow until the underlying condition changes.
Maintaining the lapsed file
Triage is not a one-time exercise. The lapsed file is alive. New names enter it every week as cadences slip; old names exit it as clients rebook or as service-history flags get added. Two operational habits keep the file usable.
First, the desk runs the triage on a 30-day cycle, not on demand. Running it monthly means the working segment stays current and the no-outreach categorizations stay fresh. Running it only when the owner remembers means the file decays and the next blast goes to a stale list.
Second, every new lapse is tagged at the moment it crosses the threshold. The CRM flags clients automatically when the gap-since-last-visit exceeds the service-category cadence multiple. The desk then reviews the new lapses in the weekly meeting and applies the fourth-cut flags (service issue, do-not-contact, one-time event) at the point the information is freshest. Waiting until the next quarterly triage to apply those flags loses signal.
For the consultation system that captures the data the triage runs on, see the consultation system that prevents the next lapse and the nine-step consultation framework.
The 30-day exercise
The fastest way to see whether the triage is real is to run it once on a quarter of the lapsed file.
Pull every client whose last visit was 12 to 28 weeks ago. Apply the four cuts in order. Count the names that survive each cut. Sort the survivors by service category, cycles past lapsed, and LTV. The exercise takes a focused afternoon for a salon with 150 names in scope, longer for a busier book.
What the exercise produces is two things. First, the actual working segment for the quarter, ready to feed the win-back sequence. Second, a clearer picture of the salon's retention pattern. The shape of the surviving segment (heavy on one service category, heavy in one stylist's book, concentrated in a particular cadence cycle) often surfaces a structural issue the salon was not seeing in the dashboard. The triage is a retention diagnostic in addition to a sorting operation.
If the working segment after triage is suspiciously large, the cuts were too generous. Run them again, tighter. If the segment is suspiciously small, the cadence cycles or the LTV thresholds are too aggressive. Calibrate against the salon's actual cadence math, not against an industry benchmark.
For the connection to no-show patterns that often precede a full lapse, see the no-show recovery sequence.
Embedded FAQ
What is the first cut in lapsed-client triage?
Service category. A lapsed single-process client at 12 weeks is a different problem from a lapsed balayage client at 12 weeks. The single-process client is two cycles past her window; the balayage client is on cadence. Mixing service categories at the top of the triage produces sequences that read as generic to every recipient, regardless of how well the rest of the segmentation is done.
How long lapsed is too long lapsed?
Past three full cadence cycles, most clients have established a new stylist relationship or a new home-maintenance routine. They are reachable but not with a standard win-back. They belong in a lower-frequency reactivation pool, typically one or two named touches per year, run sparingly and never inside the same sequence that targets the one-to-two-cycle segment.
Should we triage by lifetime value or by recency?
Both, in that order. First by service category and cadence cycles, then by lifetime value within each segment. Outreach runs first against the highest-value lapses inside one to two cadence cycles, where the recovery probability is highest. LTV alone misses recent low-value lapses worth touching; recency alone misses the high-value lapses that warrant the most effort.
Does the reason for the lapse matter?
Yes, but the salon usually does not know it reliably. The triage assumes ignorance about the cause and treats clients differently based on what the salon can observe: service category, lifetime value, recency, and documented service notes. Asking clients directly during the win-back sequence rarely produces honest answers; clients tend to give the answer that is easiest to send rather than the answer that is true.
Who should not be in the win-back list at all?
Clients with documented service issues that were not resolved; clients who explicitly asked not to be contacted; chronic no-shows already moved to deposit-only booking; clients whose last visit was a one-time event (wedding, photoshoot, traveling guest); and clients whose contact information has bounced. These five categories are removed entirely at the fourth cut and do not appear in any lapse-triggered workflow until the underlying condition changes.
How often should the triage run?
On a 30-day cycle, not on demand. Monthly triage keeps the working segment current and the no-outreach categorizations fresh. New lapses are flagged automatically when the gap-since-last-visit exceeds the service-category cadence multiple, and the desk reviews them in the weekly meeting to apply fourth-cut flags while the information is still recent.
Conclusion
Run the four cuts in order, calibrate against the salon's actual cadence math, and accept that the working segment should feel uncomfortably small. The 60 to 150 names that survive triage are the salon's retention opportunity. The 450 names that did not survive were costing the salon signal without ever being a real opportunity. Triage is the precondition for the win-back sequence, not a step inside it.
CTA
Hand the triage and outreach to a system designed for it. See how the AI front desk segments your book and runs the cadence-aware lapse workflow on a 30-day cycle.
