Designs, supplies and fits kitchens and bathrooms for homeowners, housebuilders and landlords across Nottinghamshire, Derbyshire and south Yorkshire. 72 staff.
Verdict: The map covers 6 stages, 27 touchpoints, 3 customer types; the whole journey owned by Ruth A, operations director, from November 2026. The moments that matter, chosen on the evidence before the last 6 months: Snag reported; Delivery and fitting dates confirmed; Delivery to site; Return visit to fix; Final invoice. Of the 64 complaints in those 6 months, held back as the baseline, 73% were on them. The same 5 moments come top by frequency alone: the severity weighting didn't decide them; 20 items re-graded blind by Priya N, installation surveyor: 80% agree exactly, 100% within one point. Moments with a fix, an owner and a due date: 5 of 5; done: 0. Today the moments give 9.34 complaints per 100 orders (give or take 2.73), against 3.38 elsewhere; on the moments the customer had to contact again in 21% (10 of 47, give or take 12 points); median days to resolve: 13 days (38 resolved of those raised at least 30 days ago; 0 still open); first invoices on time and right first time: 70% (139 of 200, give or take 6 points). Direction only: about 33 complaints on the moments are expected by then, so the smallest fall it could show is 68.4%, more than the 30% a fix could realistically give. Measure again on 16 Apr 2027, the same way and net of complaints elsewhere, with an early read on 14 Jan 2027. Not yet measured: fill in the Value Case after 16 Apr 2027.
Bar: the severity total at each stage (each customer item counts once; 1 is an irritation, 3 stopped them using it or made them say they would go elsewhere). Red: the share from the moments that matter. “First problem” has no owner: nobody is responsible for it from start to finish, and it has the second-highest severity total. Staff observations are listed but not counted.
| Moment | What the customer expects | What happens now | Fix and owner | Status |
|---|---|---|---|---|
| 1. Snag reported First problem · score 49 from 25 pains used to choose (4 kinds of source) · 2 by frequency alone · 8 baseline complaints, £140 remedy each | A reference and a date for the return visit on the first call | Taken by whoever answers; no single log; customers chase for a date | One snag log per job; a return date given on the first call; customer care owns it until the customer signs it off Sam K, customer care · by 11 Dec 2026 | Not started |
| 2. Delivery and fitting dates confirmed Order confirmed · score 47 from 29 pains used to choose (4 kinds of source) · 1 by frequency alone · 13 baseline complaints, £70 remedy each | Dates in writing when the order is confirmed, enough notice to plan around (three weeks for tenants), and no change without a call and a reason | Dates given by phone; moved by the planner when a team is short, often with a few days' notice | Dates confirmed in writing at order with a named planner; any change agreed by phone and in writing at least ten working days ahead, or three weeks ahead where tenants need notice Kim R, installation planner · by 30 Nov 2026 | In progress |
| 3. Delivery to site Delivered or installed · score 34 from 16 pains used to choose (4 kinds of source) · 3 by frequency alone · 15 baseline complaints, £430 remedy each | Everything arrives together, undamaged, before the fitters start | Picked the afternoon before delivery; shortages and damage found by the fitters on day one | Every order picked and checked against the order list two working days before delivery; shortages ordered before the fitters arrive Tony B, warehouse manager · by 13 Nov 2026 | In progress |
| 4. Return visit to fix First problem · score 25 from 12 pains used to choose (3 kinds of source) · 5 by frequency alone · 3 baseline complaints, £520 remedy each | Fixed in one visit, with the part on the van | Sent without checking the part is in stock; a second visit is often needed | Parts checked and on the van before a return visit is booked; the visit is booked only when they are Lee W, installation manager · by 4 Dec 2026 | Not started |
| 5. Final invoice First invoice · score 20 from 12 pains used to choose (4 kinds of source) · 4 by frequency alone · 8 baseline complaints, £60 remedy each | One invoice, after the snags are fixed, matching the quotation and the extras agreed | Raised from the job system's completion date, before snag sign-off; extras sometimes missing their agreement date | Final invoice raised only after the snag list is signed off; each extra listed with the date it was agreed Mags T, accounts · by 20 Nov 2026 | In progress |
Chosen on the evidence before the last 6 months; those months’ complaints are held back as the baseline, so the moments aren’t measured on the complaints that picked them. The same 5 moments come top by frequency alone: the severity weighting didn't decide them. 20 items re-graded blind by Priya N, installation surveyor: 80% agree exactly, 100% within one point. The expectation comes from customer evidence (interviews, reviews and complaints), never the staff workshop. Next in line, for the next round once these are done: fitter arrival on day one; daily tidy and protection; trades sequence (plumber, electrician, tiler).
| Source | Items | Pains | Praise |
|---|---|---|---|
| Complaints log | 131 | 131 | 0 |
| Online reviews | 34 | 14 | 20 |
| Call and email log | 40 | 40 | 0 |
| Customer interviews | 24 | 19 | 5 |
| Staff workshop (not counted) | 8 | – | – |
References are order, review or interview numbers; no names. One person graded every item against the written scale (a customer who had to contact again is at least 2) and will grade the re-measure the same way. 20 items re-graded blind by Priya N, installation surveyor: 80% agree exactly, 100% within one point.
| Customer type | Orders | Complaints | Per 100 orders | Worst touchpoint for them |
|---|---|---|---|---|
| Homeowners | 612 | 79 | 12.91 | Delivery to site |
| Housebuilders | 238 | 33 | 13.87 | Snag reported |
| Landlords and housing associations | 146 | 19 | 13.01 | Delivery and fitting dates confirmed |
Each type gets its own one-page map in the handover document: the stages, the touchpoints they meet, what they expect and the worst point for them.
| Measure | Now | From |
|---|---|---|
| Complaints per 100 orders on the moments (main) | 9.34 | 47 complaints on 503 orders in the last 6 months; give or take 2.73 |
| Complaints per 100 orders elsewhere (the background) | 3.38 | 17 complaints; the re-measure nets off any change here |
| Customer had to contact again | 21% | 10 of 47, give or take 12 points |
| Median days to resolve | 13 | complaints raised at least 30 days ago: 38 resolved, 0 still open |
| First invoices on time and right first time | 70% | 139 of 200 within 5 days of snag sign-off, give or take 6 points |
Direction only: about 33 complaints on the moments are expected by then, so the smallest fall it could show is 68.4%, more than the 30% a fix could realistically give. It is worked out the same way as the net change at the re-measure, from complaints on the moments and elsewhere, before and after a realistic fall. So the main measure is read for direction, beside the two per-case measures (contacted again, days to resolve) and the invoice check, which every order feeds. The re-measure uses the same complaints log, the same definition and the same five touchpoints, and nets off any change in complaints elsewhere in the journey. Not yet measured: fill in the Value Case after 16 Apr 2027.
For scale only: a complaint on the moments costs about £284 on average to handle and put right (handling hours assumed; each moment’s remedy cost from the records, from £60 to £520). If complaints on the moments fell to the owner’s target of 6.00 per 100 orders (an assumption), that would be worth £9,548 a year; a realistic fall of 30% would be worth £8,004. Neither is a forecast, and customers kept are not counted. Fee and team time: £5,880.
Evidence over 12 months: complaints from the complaints log kept by customer care in a shared spreadsheet (100% of the 131 in the log tagged), with reviews, call and email reasons and 8 interviews; orders from the order system: orders fitted or delivered in the period, by customer type. A complaint is: any expression of dissatisfaction about an order, by phone, email, letter, in person or in reply to a review, entered in the complaints log against the order number. Each item counts once per reference and touchpoint; staff observations are kept out of the counts. Severity is graded 0–3 against written words; 20 items re-graded blind by Priya N, installation surveyor: 80% agree exactly, 100% within one point. The ranking is frequency × severity, nothing else, and the same 5 moments come top by frequency alone: the severity weighting didn't decide them. The moments were chosen on the evidence before the last 6 months, and the baseline counts only those months' complaints, so the complaints that picked the moments aren't the ones measured; some regression to the mean can remain (the moments may have been unlucky in both halves), and netting against complaints elsewhere guards against what moves everywhere. Direction only: about 33 complaints on the moments are expected by then, so the smallest fall it could show is 68.4%, more than the 30% a fix could realistically give. The target of 6.0 per 100 orders, the 2.5 hours to handle a complaint and the 30% realistic fall are assumptions; the remedy cost per moment (an average of £284 a complaint with handling) is from the records. Only the cost of handling and putting right complaints is counted, not customers kept or won back.
Fictional, continuing the example. Sallowfield chose the guided route and, in the final session, set up an agent in an AI assistant the business already uses, connected read-only to the complaints log and the order system (its IT confirmed this before it committed). Early on the first Monday of each month it counts the complaints customer care has tagged over the last 3 months against each touchpoint, sets each moment that matters against its baseline rate, and sends the journey owner a short note for the monthly review. It flags a touchpoint outside the moments with 5 or more complaints in the window, and a moment with that many whose rate is 25% or more above its baseline: the journey owner's settings on Inputs. It never tags, grades or answers a complaint, and never changes the map. Nothing left the business, and no decision changed, without one of its own people approving it.
| Date | What the agent flagged | What was decided |
|---|---|---|
| 16 Oct Flag | Test run at set-up: Delivery to site was rising, with 10 complaints in three months (3.97 per 100 orders against 2.98 in the baseline months). | Ruth A left it as it was: the pick-and-check fix was already in progress, due 13 November. |
| 2 Nov Flag | Three October complaints in the log had no touchpoint, so they were listed and left out of the counts. | Sam K tagged them the same morning (two to Delivery to site, one to Snag reported). |
| 7 Dec Flag | A new hot spot outside the five moments: the trades sequence, with 6 complaints in three months (2 at set-up). | Ruth A added it to next in line, and asked Kim R to book the electrician with the fitting team from January. |
| 7 Dec Flag | The final-invoice fix was past its 20 November date and still in progress. | Ruth A agreed a new date of 18 December with Mags T, once the job system has a sign-off date field. |
| 4 Jan No flag | Delivery to site was back below its baseline: 2.03 per 100 orders, with the pick-and-check fix done since 13 November. | Nothing needed. |
| 4 Jan Flag | Snag reported was rising: 7 complaints in three months, 2.85 per 100 orders against 1.59. Its fix had not started. | Sam K to give a return date on the first call from 11 January; checked at the February run. |
At the check-in on 30 November, after the first monthly run (about day 45), Geoff and Ruth A read the log together; no rule needed changing. The agent added two things: the month's numbers were ready for the first five minutes of every monthly review, and a new hot spot (the trades sequence) and a stalled fix (the snag log) were raised within a month, not at the re-measure. Its readings are a running count only: at these volumes they move by chance, so the early read on 14 Jan 2027 and the re-measure on 16 Apr 2027 are still taken by people, the same way, and judge the fixes. The months after add nothing to the figures above, and no £ figure is claimed for them.