Customer Interview Programme · October 2026

Mellstow Chemicals Ltd

Distributes cleaning, process and water treatment chemicals to food manufacturers, metal finishers and water treatment contractors from a warehouse and technical office in Wednesbury. 58 staff.

Worked example. Mellstow Chemicals and its figures are fictional. All figures come from the worked-example spreadsheet: ALL CHECKS PASSED (with 5 notes).
The headline

30 customers interviewed; 11 themes, every one with an owner

Customers interviewed
30
S1 11 of 10; S2 10 of 10; S3 9 of 10 (31 interviews)
Themes
11
each from at least 3 customers; 2 signals not yet themes
Quotations cleared
31
for internal use; marketing use asked for separately
Catalogue uses ready now
4 of 7
6 products; 3 waiting, all for marketing permission

Verdict: 30 customers across 3 segments (S3 9 of 10, short of target) gave 11 themes, each from at least 3 customers, and 2 signals not yet themes. The top reason customers buy: technical advice on the phone (12 of 30); the top reason they nearly left: missed or split deliveries (8); the top thing they say they would pay more for: guaranteed delivery slot (6; stated, not tested). 2 themes rest mostly on interviewers with a stake in the answer: read with care. 31 quotations are cleared for internal use and 0 for marketing use. Every theme has an owner and an action. It feeds 6 catalogue products: 4 of 7 uses are ready now; first waiting: Value Proposition and Messaging: waiting for at least 3 quotations cleared for marketing use, asked for separately after the research (now 0). Still turning up new themes in the last 3 interviews: S3; one or two more interviews there before its themes are settled.

Who we heard from

Spread by value and tenure, with at-risk and lapsed customers included

SegmentCustomers (of target)Value: high / middle / lowerTenure: new / established / long-standingAt risk or lapsedInterviews by someone with a stakeStatus
S1 Food and drink manufacturers
Hygiene or technical manager
11 of 10
11 interviews
3 / 5 / 32 / 5 / 431On target
S2 Metal finishing and engineering firms
Production or purchasing manager
10 of 10
10 interviews
2 / 5 / 32 / 4 / 432On target
S3 Water treatment contractors
Owner or operations manager
9 of 10
10 interviews
2 / 4 / 32 / 4 / 321Short of target
1 new theme in the last 3 interviews

46 customers invited (47 contacts): 32 accepted, 7 declined, 8 no reply. 4 of 31 interviews by someone with a stake (2 by the customer's own account manager); 0 at-risk or lapsed customers heard by someone with a stake. 25 interviews recorded with consent, 6 notes only; 0 recordings kept past 30 days; notes and Findings (pseudonymised) deleted or stripped of references by 9 Apr 2028. £1,014,800 of £6,150,000 sold to these segments in the last 12 months (17%). Every at-risk and lapsed customer was interviewed by the consultant, who has no stake in why they were unhappy. The managing director interviewed two customers, and two metal-finishing customers asked for their own account manager; those four interviews are flagged, and so is any finding heard by the person responsible for that area. One customer had two contacts interviewed: two interviews, one customer.

What customers said

89 statements from 30 customers: 45 positive, 23 negative, 1 mixed, 20 neutral

Why they buyTechnical advice on the phone12 customersNext-day delivery we can rely on8 customersSafety paperwork right first time6 customersNearly leftMissed or split deliveries8 customersPrice rises with no explanation7 customersAccount manager changed without notice2 customers · signal, not yet a themePay more for (stated)Guaranteed delivery slot6 customersSite visits and dosing checks5 customersOthers consideredNational distributors on price9 customersBuying direct from the manufacturer2 customers · signal, not yet a themeWant nextSmaller pack sizes7 customersOnline ordering and order tracking6 customersWorking with usOne person who knows our site7 customers

Mostly positive Mostly negative Neutral Signal: fewer than 3 customers

Each bar counts distinct customers, not statements: a customer who came back to the same point twice counts once. A theme needs at least 3 customers; below that it is a signal, listened for in the next round but not acted on yet. 12 statements coded again by a second person; 11 agreed (92%).

The themes

Each theme in the customer's words, by segment (S1 / S2 / S3)

AreaTheme and what countsCustomersS1 / S2 / S3Heard without a stakeA quotation cleared for internal use
Why they buyTechnical advice on the phone
They value ringing and getting a straight answer from someone who knows the chemistry: dilution rates, compatibility, what to use instead.
125 / 4 / 39“When we changed the cleaning cycle on line two, I rang and had a chemist on the phone in five minutes. Nobody else does that.” (S1)
Why they buyNext-day delivery we can rely on
Orders placed by the cut-off arrive the next working day, and they plan around it.
84 / 3 / 16“Order by three and it's here next morning. We plan the hygiene shifts round that.” (S1)
Why they buySafety paperwork right first time
Safety data sheets, COSHH information and certificates of analysis come with the order and are right, so audits go smoothly.
64 / 0 / 22 stake“If you went, the thing I'd miss most is the paperwork: the safety data sheets are always right and the auditor never queries them.” (S1)
Nearly leftMissed or split deliveries
An order arrived late, short or in more than one drop, and it cost them time or production.
82 / 3 / 37“A short delivery of degreaser stopped the line for half a day. Nobody rang to warn us.” (S2)
Nearly leftPrice rises with no explanation
A price rise arrived without notice or a reason, and they looked elsewhere or thought about it.
73 / 3 / 16“The last price rise just appeared on the invoice. No letter, no reason. That's when we started getting other prices.” (S1)
Pay more for (stated)Guaranteed delivery slot
They say they would pay extra to know the delivery arrives in a set slot or on a set day.
64 / 2 / 03“A guaranteed morning slot would be worth paying for. A late delivery means a cleaning shift stood around.” (S1)
Pay more for (stated)Site visits and dosing checks
They say they would pay for a visit to check dosing, test water or train their staff on site.
50 / 0 / 54“If you'd come out and check dosing on our bigger sites, we'd pay for that. We give that advice away now and so do you.” (S3)
Others consideredNational distributors on price
They compare us with national distributors, usually on price.
93 / 4 / 27“Every quarter I get the national firm's price list and check your top five lines against it.” (S1)
Want nextSmaller pack sizes
They want products in smaller packs than we sell.
73 / 2 / 27“We use a 25-litre drum over three months and half of it goes out of date. A 5-litre would suit us better.” (S1)
Want nextOnline ordering and order tracking
They want to order, reorder or track deliveries online.
62 / 2 / 22 stake“If I could order from my phone on site, I'd order more from you.” (S3)
Working with usOne person who knows our site
They value dealing with one person who knows their site and what they buy.
72 / 3 / 24“One person who knows every site we've got. That's worth a lot.” (S3)

"Pay more for" is stated intent: what customers say they would pay, not what they would. The Price Rise Plan tests it against the volume each line can afford to lose. Heard mostly by someone with a stake (the managing director, the customer's own account manager or the person responsible for that area): safety paperwork right first time, online ordering and order tracking. Read them with care.

Surprises

What the owner didn't expect

The biggest surprise: smaller customers want smaller packs, and water treatment contractors would pay for the site visits and dosing advice Mellstow gives away. Signals to listen for next time: account manager changed without notice, buying direct from the manufacturer.

Actions

12 actions; 11 of 11 themes with an action; 4 owners

ThemeActionOwnerDueFeeds
Technical advice on the phonePut the technical line and the chemists at the centre of the website and proposals; keep the line staffed 8am to 5pmManaging director30 Nov 2026Value Proposition and Messaging
Next-day delivery we can rely onPublish the order cut-off and report on-time, in-full delivery weeklyOperations director15 Dec 2026None (internal fix)
Safety paperwork right first timeAdd a sample safety paperwork pack to every food proposalCustomer service manager15 Dec 2026Proposal Rewrite
Safety paperwork right first timeAfter the findings are shared, contact the food customers with strong audit records separately to ask about a case studyManaging director31 Jan 2027Customer Case Study Programme
Price rises with no explanationGive 60 days' notice of the next rise, with a letter explaining the cost behind itManaging director31 Jan 2027Price Rise Plan
Missed or split deliveriesFlag any customer with two missed or split deliveries in a quarter as at risk and ring themOperations director30 Nov 2026Customer Retention Review
Guaranteed delivery slotPrice a timed-delivery option for food and metal finishing customersOperations director28 Feb 2027Price Rise Plan
Site visits and dosing checksOffer paid site visits and dosing checks to water treatment contractorsAccount manager (water treatment)28 Feb 2027Key Account Growth Plan
National distributors on priceAnswer price comparisons with the advice and paperwork, not a discountManaging director31 Dec 2026Value Proposition and Messaging
Smaller pack sizesTrial 5-litre packs of the six fastest-selling lines for smaller accountsOperations director31 Mar 2027Key Account Growth Plan
Online ordering and order trackingScope delivery-tracking emails from the existing order systemCustomer service manager31 Jan 2027None (internal fix)
One person who knows our siteTell customers before an account manager changes, and introduce the new one on siteManaging director30 Nov 2026Customer Retention Review
How the value is shown

Unlocks: the catalogue products the findings feed

Catalogue productWhat it needsStatus
Value Proposition and MessagingReasons they buy in their wordsReady now
Value Proposition and MessagingQuotations cleared for marketing useWaiting for at least 3 quotations cleared for marketing use, asked for separately after the research (now 0)
Price Rise PlanWhat they say they'd pay more for (stated)Ready now
Customer Retention ReviewWhy customers nearly leftReady now
Key Account Growth PlanUnmet needs and other suppliersReady now
Proposal RewriteQuotations cleared for marketing useWaiting for at least 3 quotations cleared for marketing use, asked for separately after the research (now 0)
Customer Case Study ProgrammeQuotations cleared for marketing useWaiting for at least 3 quotations cleared for marketing use, asked for separately after the research (now 0)

31 cleared for internal use; 0 cleared for marketing use (asked separately); 5 held without permission. Quotations go outside the business (website, proposals, case studies) only with the customer's separate permission for marketing use, asked after the research and never in the interview; none has been asked yet, so those uses wait.

No measure-later line. With 9 customers interviewed in the smallest segment, a share has a give or take of ±33 percentage points, so a re-run couldn't show a realistic change; the products it feeds measure their own results.

Time and cost: 76 team hours (£2,891 of time); fee £5,400 ex VAT.

The interview guide

15 open questions for each version, checked for closed openers, leading or assumptive wording and two asks in one

#Asked ofTopicQuestion
1AllOpeningTell me about your role: what is your part of the business trying to get done this year?
2AllWhy they chose usTell me how you first came to buy from us: what was going on at the time?
3AllWhy they chose usWhat made you choose us rather than the other options you had?
4AllWho else they consideredWho else have you looked at alongside us?
5AllWorking with usWalk me through the last time you ordered from us.
6Current customersWhat they'd loseIf we stopped trading tomorrow, what would you lose?
7Current customersWhat nearly made them leaveTell me about any time you've thought about moving some or all of your business elsewhere: what was going on?
8Current customersWhat nearly made them leaveWhat would make you more likely to move business away from us over the next year?
9Lapsed customersWhat they'd loseWhat, if anything, do you miss since you stopped buying from us?
10Lapsed customersWhat nearly made them leaveWalk me through what led you to stop buying from us: what was going on?
11Lapsed customersWhat nearly made them leaveWhat could we have done that would have kept you?
12AllWhat they'd pay more forWhat, if anything, would be worth paying more for?
13AllWhat they'd want nextWhat do you wish we did that we don't do today?
14AllWhat they'd pay more forHow do you judge whether what you pay us is fair for what you get?
15AllHow they describe usWhen you describe us to a colleague, what do you say?
16AllChanges comingWhat's changing in your business over the next year or two that we should know about?
17AllCloseIf you were running our business, what's the one thing you'd change first?
18AllCloseWhat haven't I asked about that you expected me to?
19AllPrice test (optional)If our prices on your main lines went up, at what rise would you start getting prices from elsewhere?

The optional price question gives a stated, hypothetical range from a handful of customers: a starting point for the Price Rise Plan, never a price.

How confident are we?

What this rests on

The findings are what 30 customers said in 31 interviews, coded by the team into themes; a theme needs at least 3 distinct customers, counts customers not statements, and shows how many said it, not how many of all customers think it. The customers were chosen across value and tenure bands and included at-risk and lapsed customers, but those who agreed to talk may differ from those who didn't, and interviewers with a stake in an answer are flagged. What customers say they would pay more for is stated intent, not tested. The target of 10 customers a segment is a judgement drawing on published saturation tests (Guest, Bunce and Johnson 2006; Hennink, Kaiser and Marconi 2017), not a guarantee. Interviews show what customers think and why; they don't measure how much it is worth: the products the findings feed do that, from the owner's records. The hours are estimates. The Findings sheet is pseudonymised personal data; UK GDPR points follow ICO guidance and are not legal advice.

The findings are finished when they are handed over. Run the programme again in 12 to 18 months, with the same guide, to hear whether the fixes landed.

The months after

What the agent prepared and flagged, and what people decided

Fictional, continuing the example; the log is on the Agent sheet of the worked-example workbook, and nothing in it changes the findings above, which stand as at 15 October 2026.

Mellstow took the guided route. Because this agent's main work is the fieldwork, it was set up at the start of it: on 17 August 2026 the client lead and Geoff set it up together, in an AI assistant the business already uses, connected read-only to the programme folder on the shared drive (the workbook and the interviewers' written notes) and able to save drafts in a shared drafts document without sending; Mellstow's IT confirmed the connections beforehand. The same day Mellstow updated its privacy notice to cover the interviews; the notes stay in its own files. During the fieldwork (the first 5 rows of the log) it drafted at most 15 invitations a week and one reminder after 7 days, each with the privacy information and nothing to sell, for the customer service manager to approve and send from her own email; it proposed interviewers without a stake, and flagged the customer who asked for its own account manager; and it drafted the coding of each written-up interview against the codebook, for the interviewer who ran it to confirm. Customers happy to be recorded said so in writing, replying to the invitation, before the session; the answer went on Sample and was confirmed at the start. Even so, the agent worked from the interviewers' written notes, and the recordings were kept only to check them. It never ran an interview, never wrote to a customer and never filled in the second coder's column. Nothing left the business, and no decision changed, without one of Mellstow's own people approving it.

DateWhat the agent prepared or flagged, and the evidenceKindApproved or decided byWhat happened
24 Aug 2026Drafted the first 15 invitations (S1 and S3 first, with every at-risk and lapsed customer on those lists), each with the privacy information from the template and nothing to sell
Sample (rows marked to invite); the invitation in the deliverable template; Inputs: 15 a week
DraftCustomer service managerShe shortened the opening paragraph and sent all 15 from her own email on 25 Aug; Invitation set to Invited on Sample
1 Sep 2026CUS-305 replied asking to be interviewed by its own account manager; the agent had proposed the customer service manager
The customer's reply in the approver's mailbox; Inputs: 0% of interviews by the customer's own account manager
FlagManaging directorGo ahead with the account manager (metal finishing), at the customer's request; flagged on Sample and on every finding as an interview with a stake
7 Sep 2026Drafted one reminder each to the 9 contacts with no reply after 7 days (no second chase)
Sample: Invited with no reply; replies in the approver's mailbox
DraftCustomer service managerSent from her own email on 7 Sep; the answers were entered on Sample as they came in
21 Sep 2026Drafted the coding of the 12 interviews held 7 to 18 Sep: one theme and a sentiment for each point, against the codebook on Themes
Each interviewer's written notes in the programme folder (worked from the notes; interviews were recorded only where the customer had agreed in writing, replying to the invitation, before the session, and the recordings were kept only to check the notes)
DraftEach interviewerFour points re-coded by the interviewers, and one new theme added to the codebook by the managing director; the second coder's column left for the blind second coder
12 Oct 2026S3 at 9 of 10 customers, and its last 3 interviews still turned up a theme not heard there before; 3 S3 reserves not yet invited
Coverage; Themes (new in S3's last interviews); Sample
FlagManaging directorKeep the findings date at 15 Oct; invite two S3 reserves after handover and code their interviews on a working copy, as the start of the next round
19 Oct 2026Listed 7 recordings reaching 30 days by 25 Oct, with no deletion date on Sample
Sample: recorded with consent, no deletion date; Inputs: 30 days
FlagCustomer service managerEach interviewer deleted their recordings; the dates entered on Sample by 23 Oct
19 Oct 2026Drafted invitations to two S3 reserves (CUS-580 and CUS-586), research only, with the privacy information
Sample (S3 reserves); the managing director's decision of 12 Oct
DraftCustomer service managerSent from her own email on 20 Oct; CUS-580 accepted for 4 Nov with the consultant; CUS-586 declined
22 Oct 2026Drafted the coding of CUS-363's interview (21 Oct, the managing director): it accepted in September, but its first free date fell after the findings date, so the managing director held it then for the next round
The managing director's written notes in the programme folder
DraftManaging directorFive points confirmed on the working copy for the next round; the findings handed over on 15 Oct are unchanged
2 Nov 2026Listed the last 10 recordings reaching 30 days by 8 Nov
Sample: recorded with consent, no deletion date
FlagCustomer service managerAll deleted and dated on Sample by 6 Nov; no recording now kept
9 Nov 2026Drafted the coding of CUS-580's interview (4 Nov, the consultant)
The consultant's written notes in Mellstow's programme folder
DraftConsultantOne point moved from Missed or split deliveries to Guaranteed delivery slot; S3 now 10 customers on the working copy, with nothing new in the interview
16 Nov 2026Test run of the next round, not sent: starts 15 Jan 2028 (15 months after the programme date), preparation from 4 Dec 2027; the notes are kept until 9 Apr 2028, so the two rounds can be compared
Inputs (the agent rows and the notes period); Sample; Summary
LoggedManaging directorRead at the 30-day check-in with Geoff; plan kept and the dates put in the diary

After handover the agent turned to the housekeeping and the next round. It listed the 17 recordings still held at handover as each reached 30 days, and the interviewers deleted every one by 6 November. S3, short of target and still turning up a new theme, got two more invitations; the one interview that followed added nothing new, so S3's themes look settled for the next round. CUS-363, an S2 customer that accepted in September but could not meet before 21 October, was interviewed then by the managing director and coded on the working copy for the next round; the findings handed over on 15 October are unchanged. At the 30-day check-in on 16 November, Geoff and the managing director read the fieldwork log together with a test run of the next round's plan (not sent): it starts on 15 January 2028, preparation begins on 4 December 2027, and the notes are kept until 9 April 2028 so the two rounds can be compared. His upkeep of the set-up ended there. Approving drafts and confirming the coding took the customer service manager about 30 minutes a week during the fieldwork and each interviewer about 10 minutes an interview (estimate). The findings, the actions and the value case are unchanged.