Replacement windows, doors and conservatories for homeowners, surveyed and fitted from three showrooms in Worcestershire. 72 staff.
Verdict: In 36 shops of the business, a prospect got 43% of the standard points on average (give or take 9 points). 50% were responded to within the time standard and 14% got no response at all; where a follow-up was due, 35% came within the standard and 48% came at all. By channel: Walk-in best (71%), Web form worst (31%). By branch: Worcester showroom best (58%), Kidderminster showroom worst (30%). Lowest points: Asked for the survey or appointment (35%); Followed up within the standard (35%); Picked up the cue the shopper dropped (39%). Rivals: Rival A 37% (6 shops) against our 35% (18 shops); Rival B 36% (6 shops) against our 35% (18 shops). Fix list: 8 fixes (8 with an owner and a due date); the three lowest points with a fix: 3 of 3; last due 20 Nov 2026. Scorers: 8 of 36 shops double-scored, blind to the first score; agreement 93%; scored on the consultant's side, so use an independent or the owner's own scorer for the re-shop. Re-shop 11 Jan 2027 to 29 Jan 2027 (moved clear of Christmas and New Year) with the same scenarios, standard and shopper brief, and score the before and after write-ups together, blind; a change under 13 points could be chance. Not yet measured: re-shop 11 Jan to 29 Jan 2027 and fill in the Value Case after 10 Feb 2027.
| Response time, our shops | Within the standard | Up to twice | Slower | No response | Shops |
|---|---|---|---|---|---|
| Phone standard: 15 seconds | 4 | 3 | 0 | 2 | 9 |
| Web form standard: 4 working hours | 2 | 1 | 4 | 2 | 9 |
| Email standard: 4 working hours | 6 | 0 | 2 | 1 | 9 |
| Walk-in standard: 2 minutes | 6 | 2 | 1 | 0 | 9 |
Walk-in visitors were greeted quickly and asked about the job. Web forms were answered slowly, if at all; emails were answered quickly, but mostly with a brochure and few questions. Two of nine phone calls rang out with no voicemail. Nine shops a channel is a small base: read the differences as direction.
| Standard point | All | Phone | Web form | Walk-in | Fixes | |
|---|---|---|---|---|---|---|
| 1. Gave the business name and their own name | 65% | 71% | 100% | 50% | 44% | 1 |
| 2. Took the shopper's name and a way to contact them | 44% | 43% | – | – | 44% | 0 |
| 3. Asked at least two questions about what the shopper needs | 42% | 43% | 14% | 12% | 89% | 1 |
| 4. Asked when the shopper wants the work done | 52% | 43% | 43% | 25% | 89% | 0 |
| 5. Picked up the cue the shopper dropped Lowest 3 | 39% | 57% | 0% | 12% | 78% | 1 |
| 6. Gave the right answer to the key question | 77% | 71% | 86% | 75% | 78% | 0 |
| 7. Explained in plain words | 68% | 57% | 57% | 62% | 89% | 0 |
| 8. Gave a specific reason to choose us | 52% | 29% | 29% | 50% | 89% | 0 |
| 9. Gave a price range or said when a quotation would come | 45% | 43% | 29% | 38% | 67% | 0 |
| 10. Offered a next step with a day or time | 45% | 43% | 43% | 12% | 78% | 0 |
| 11. Asked for the survey or appointment Lowest 1 | 35% | 43% | 14% | 25% | 56% | 1 |
| Timed: Responded within the time standard | 50% | 44% | 22% | 67% | 67% | 3 |
| Timed: Followed up within the standard Lowest 2 | 35% | 29% | 43% | 12% | 56% | 1 |
| Branch | Shops | Average score | Responded in time | Followed up in time |
|---|---|---|---|---|
| Worcester showroom | 12 | 58% | 67% | 64% |
| Kidderminster showroom | 12 | 30% | 42% | 27% |
| Bromsgrove showroom Fewer than three people: results may identify someone | 12 | 41% | 42% | 11% |
Worcester showroom best (58%), Kidderminster showroom worst (30%). Results are by branch, never by person; Bromsgrove has two people handling enquiries, so its figures could identify them, and are shared with them on that basis.
Best example Walk-in, bi-fold doors · Worcester showroom · shop score 100%
Greeted within a minute by name and role. Asked five questions about the opening, the floor level and how the garden is used; asked about the neighbour's leaking doors and showed the drainage track on a display door. Gave the 6-8 week lead time, offered a survey on Tuesday at 10 or Thursday at 2 and asked to book it. Called next day to confirm.
Worst example Web form, replacement windows · Kidderminster showroom · shop score 0%
Web form sent at 10.20 on a weekday morning. No reply, no call and no email in 8 working days. The form's confirmation page said "we'll be in touch within 24 hours".
Worst example Phone, composite front door · Kidderminster showroom · shop score 15%
Answered after 27 seconds (about nine rings) with "Hello, showroom". Gave a price range when asked but no questions about the house or the door; didn't take up the mother and the walking frame. Ended "have a look on the website and give us a ring back". No name taken.
Staff are referred to by role only; the shoppers wrote notes, nothing was recorded.
| Rival (never named) | Shops | Their score | Responded in time | Followed up in time | Ours on the same (shops) |
|---|---|---|---|---|---|
| Rival A | 6 | 37% | 33% | 40% | 35% (18) |
| Rival B | 6 | 36% | 33% | 0% | 35% (18) |
Rivals were asked the same questions as an ordinary enquiry, with no surveys or visits booked, and their staff aren't named; once their follow-up was recorded the shoppers closed the loop (“we’ve decided to go elsewhere”) and unsubscribed. On like-for-like phone and web-form shops the business is no better than either rival: the gap is a few points, well inside the give or take. They are shopped again at the re-shop as the background: they get none of the fixes, so a change in their scores shows how much the shoppers, scorers and season moved. With six shops each, that background has a wide give or take of its own, shown beside it on the Value Case.
| Problem found | The fix | Owner | Due |
|---|---|---|---|
| Advisers don't ask the shopper to book the survey Asked for the survey or appointment · All showrooms | End every enquiry with two survey times; practise it in the Monday huddle | Claire B | 20 Nov |
| Nobody follows up a quotation request or a visit Followed up within the standard · All showrooms | Every enquiry gets a follow-up task for two working days later in the CRM | Claire B | 13 Nov |
| Cues about the household go unasked Picked up the cue the shopper dropped · All showrooms | A five-question enquiry sheet beside every phone and in the reply template, ending "anything else we should know?" | Tom H | 20 Nov |
| Written replies send a brochure but no questions Asked at least two questions about what the shopper needs · Web form and email | New reply template: thanks, two questions, two survey times, the adviser's direct line | Rachel M | 13 Nov |
| Showroom phones ring out while advisers are with visitors Responded within the time standard · Kidderminster and Bromsgrove | Calls not answered in five rings go to the sales office; a missed-call list is cleared by 10am and 2pm | Rachel M | 6 Nov |
| Web forms and emails sit unread until the end of the day Responded within the time standard · All showrooms | Forms and emails go to one shared inbox checked on the hour; first reply is a call, not a brochure | Rachel M | 6 Nov |
| The form's confirmation page promises a reply within 24 hours Responded within the time standard · Website | Change it to "we'll call you within four working hours" once the shared inbox is live | Rachel M | 13 Nov |
| Walk-in visitors aren't told who they're talking to Gave the business name and their own name · All showrooms | Name badges; greet with name and role | Asif R | 30 Oct |
The main measure is the average shop score (43%, give or take 9 points), with the share responded to within the standard (50%) and followed up within the standard (35%). At the re-shop the same four shoppers use the same three scenarios and brief on the same channels and branches; the before and after write-ups are scored together, mixed, with the dates hidden, by the same two people, and each channel is weighted to today's mix. Blind scoring is only partial: once replies follow the new template, a scorer can tell an after shop from its wording. 8 of 36 shops double-scored, blind to the first score; agreement 93%; scored on the consultant's side, so use an independent or the owner's own scorer for the re-shop. With 36 shops again, a change under 13 points could be chance; the owner's target of 80% is well outside that. The window was moved clear of Christmas and New Year, and every fix is due before it starts. Not yet measured: re-shop 11 Jan to 29 Jan 2027 and fill in the Value Case after 10 Feb 2027.
Illustration only, not a forecast: enquiry-to-order conversion is 17.0% (109 orders from 640 enquiries over 3 months; give or take 3.0 points). If it reached the owner's assumed 19%, that would be worth £103,704 a year in gross profit (£51,852 if it moved half as far); as an illustration of scale, one point of conversion is worth £52,675. Conversion is re-read from the records on 3 May 2027 and compared with the same months a year earlier, but with about 640 enquiries a change under 4.2 points could be chance, and the owner's target is inside that: the re-read shows direction only, and the shop scores can't prove what moved it. Cost in full: £6,720 cash (fee and shoppers) and £1,120 of team time.
Fictional and illustrative. This continues the worked example from the baseline shops in September 2026 to the end of the January 2027 re-shop. Every name, figure and event is invented.
Mardlecote set up the agent before the first shop, in its own accounts: its AI assistant, the shared folder where the shoppers save their write-ups, and the CRM, read-only. The consultant approved and sent every schedule, brief and reminder to the shoppers; the showroom managers never saw a shop's date or time. The shoppers made every shop and two people scored every judged point from the notes; the agent filled only the response, the times and the follow-up from the shoppers' logs, for the scorers to check. The rules are on Inputs: a missing write-up flagged after 2 hours, "no follow-up" only after 8 working days, fix reminders 7 days ahead, and the re-shop prepared 3 weeks ahead. Nothing left the business, and no decision changed, without one of Mardlecote's people approving it.
| Date | What the agent prepared, collated or flagged | Kind | Who decided | What happened |
|---|---|---|---|---|
| 11 Sep 2026 | Shop schedule drafted: 36 of our shops (3 showrooms, 4 channels, 3 scenarios) and 12 rival shops, spread over 15 working days and 4 shoppers Teams, Scenarios and Shoppers sheets; the shopper brief | Draft | The consultant Approved with changes | The consultant swapped two walk-ins so no shopper visited the same showroom twice, then sent each shopper their own list. |
| 17 Sep 2026 | MS-025 (a phone shop by SH3): write-up not in the folder 2 hours after the shop The schedule; the shared write-up folder; Inputs (2 hours) | Flag | The consultant Decided | SH3 had saved it in the wrong folder; moved the same day. No other write-up was late. |
| 21 Sep 2026 | MS-038: Rival A's follow-up recorded on a bi-fold doors phone shop. Close-the-loop reply drafted for the shopper Shops sheet (rival row, follow-up Yes); the shopper brief | Draft | The consultant Approved | The shopper replied "thank you, we've decided to go elsewhere" and unsubscribed on 22 Sep. |
| 6 Oct 2026 | 8 of our 36 shops (22%) picked for the second scorer; their write-ups sent without the first score Shops sheet; Inputs (20% double-scored) | Logged | The consultant Approved | Double-scored by 9 Oct. Agreement 93%. |
| 14 Oct 2026 | Follow-up wait of 8 working days passed for the last shops: Not yet changed to Yes or No from the shoppers' logs, for the scorers to check The shoppers' logs; Inputs (8 working days) | Logged | Hannah W Approved | No shop left Not yet. Baseline recorded on 16 Oct. |
| 19 Oct 2026 | 31 of the shoppers' invented enquiries found in the CRM, listed for removal The CRM, read-only: the shoppers' review phone numbers and email addresses | Flag | Rachel M Decided | The sales office removed them on 20 Oct; none had a quotation raised. |
| 2 Nov 2026 | Fix 7 (name badges, due 30 Oct) not marked done Fix List (due by, done on); Inputs (7 days' notice) | Flag | Rachel M Decided | Asif R had the badges in use from 28 Oct; Done on entered. |
| 23 Nov 2026 | Fix 5 (the five-question enquiry sheet, due 20 Nov) not marked done Fix List (due by, done on) | Flag | Rachel M Decided | Tom H had the sheets beside every phone on 24 Nov. All 8 fixes done before the 4 Dec date. |
| 21 Dec 2026 | Re-shop prepared for 11 to 29 Jan 2027 (moved clear of Christmas): the same scenarios, standard, brief and 4 shoppers; 36 of our shops and 12 rival shops. Flag: both scorers are on the consultant's side Value Case (window, re-measure 10 Feb 2027); Inputs (scorer independent: No) | Draft | Rachel M Approved with changes | Rachel asked for an independent scorer for the blind scoring; the consultant arranged one. Schedules went to the shoppers on 4 Jan. |
| 29 Jan 2027 | Re-shop closed: 48 shops made, 48 write-ups in. Blind set prepared: our 72 write-ups, before and after mixed, dates hidden Shops sheet; the shared write-up folder | Logged | The consultant Approved | Scoring booked for 11 Feb, after the follow-up wait, with the independent scorer and Hannah W. |
While the baseline shops ran, one write-up was late (saved in the wrong folder) and one rival's follow-up was closed off the next working day, as the shopper brief says. After handover the notes went to Rachel M. Two fixes slipped by a few days and were done within the week; all 8 were done by 24 November 2026. On 21 December the agent prepared the re-shop: the same scenarios, standard, brief and shoppers, moved clear of Christmas to 11 to 29 January, and flagged that both scorers were on the consultant's side. Rachel asked for an independent scorer, as the review had advised.
The re-shop closed on 29 January. The before and after write-ups (72 of ours, mixed, dates hidden) are scored together after the follow-up wait ends on 10 February, so this example stops before the result: nobody yet knows whether the average shop score has moved, and a change under 13 points could be chance. The coordinator spent about 20 minutes a working day while shops ran, and Rachel about 10 minutes a week after handover (estimate).
The baseline rests on 36 of our shops by 4 independent shoppers, scored from written notes against 13 points the owner agreed; the average shop score has a give or take of 9 points, and a branch or channel has only a handful of shops, so read those as direction. Response times are counted by the shoppers in seconds, working hours and minutes; a follow-up counts as No only after 8 working days, and a shop with no response at all counts once, against the response. 8 shops were double-scored (agreement 93%). Blind scoring is only partial once replies follow a new template, because a scorer can tell an after shop from its wording. The rivals were shopped on the same scenarios and channels (12 shops) and are re-shopped as the background, because they get none of the fixes. At the re-shop each channel is weighted to today's mix and the before and after write-ups are scored together, blind, by the same people. The target score and the target conversion are the owner's assumptions; enquiries, conversion (109 orders from 640 enquiries) and gross profit per order come from the records, and any £ is an illustration, not a forecast. The re-shop shows whether the measures moved, not proof that the fixes moved them.