Social Selling Programme · October 2026

Ellerholt Building Services Engineers Ltd

Mechanical, electrical and public health engineering consultancy designing heating, cooling, ventilation, power, lighting and water systems for hospitals, universities, schools and commercial buildings, from offices in Leeds, Manchester and Sheffield. 240 staff.

Worked example. Ellerholt Building Services Engineers and its figures are fictional. All figures come from the worked-example spreadsheet: ALL CHECKS PASSED (with 6 notes).
The headline

Twelve sellers, a routine that fits, and a baseline sized before it was counted

Profiles to the standard
12 of 12
2.6 → 10.0 items of 10
Routine a week
1.7 h
48 requests a week between them; none over the time agreed
Conversations per request now
2.9%
11 of 384 requests in an 8-week tally (the main measure)
Policy agreed in time
11 of 11
the five covering the tally before it began; contact ownership reviewed by the client's solicitor

Verdict: 12 sellers in two waves each have a profile to the standard (12 of 12 meet it), 60 target accounts of their own with enough people to ask until the re-measure, and a weekly routine of about 1.7 hours that fits the time each agreed; 13 approved items to share over 13 weeks; 11 of 11 policy items agreed by go-live, those covering the tally before it began, with no tool that automates or scrapes. Measure later: conversations per request 2.9% now (main): by the re-measure a real change of the target's size would be seen four times in five; acceptance 27.1% (process), a real change of the target's size would be seen four times in five; meetings direction only; the early read nets wave 1 against wave 2 (direction only). Different times of year, and the tally ran through the summer holidays. The re-measure isn't netted, so read it with the early read's netted change. To cover the cost in full takes about 19 extra social-sourced meetings a year, 82% more than today's 23; by the re-measure a rise under 311% could be chance, so a rise that just pays its way couldn't be told from chance. Illustration only: £87,400 a year from 23 extra meetings (capped at the owner's meetings target; 41 if the main target were reached in full), not a forecast. Unlocks: Target Account List: 12 of 12 Tier 1 accounts have 3 or more people connected; Buying Group Map: 48 of 60 accounts have someone connected in two or more roles. Early read 8 Mar 2027; measure again 7 Jun 2027.

The sellers

Who is in it, their routine, their profile and their baseline

SellerWaveRequests a weekRoutine minutes (needed of agreed)Profile items (before → after)Tally: conversations of requestsAcceptedAccountsPeople left for requests needed
Idris
Business development manager, health
15104 of 1203 → 101 of 40115120 for 110
Morwenna
Business development manager, education
25104 of 1202 → 101 of 40105120 for 110
Folake
Business development manager, commercial
15104 of 1203 → 102 of 40125120 for 110
Sorcha
Associate director, mechanical (health)
24101 of 1102 → 101 of 329594 for 88
Esme
Associate director, electrical (health)
14101 of 1103 → 101 of 328594 for 88
Kwame
Associate director, mechanical (education)
24101 of 1104 → 100 of 328594 for 88
Bartek
Associate director, public health (commercial)
14101 of 1102 → 101 of 329594 for 88
Elspeth
Associate director, sustainability
24101 of 1103 → 101 of 328594 for 88
Anjali
Associate director, electrical (commercial)
14101 of 1102 → 101 of 329594 for 88
Rhodri
Regional director, Leeds
2398 of 1053 → 101 of 247571 for 66
Joel
Regional director, Manchester
1398 of 1051 → 101 of 246571 for 66
Zara
Regional director, Sheffield
2398 of 1053 → 100 of 247571 for 66
28%Idris25%Morwenna30%Folake28%Sorcha25%Esme25%Kwame28%Bartek25%Elspeth28%Anjali29%Rhodri25%Joel29%ZaraWave 1Wave 2 (the early read's comparison)Line: all sellers, 27.1%

Each seller's acceptance in the tally (the process measure), asking the way they do now. The last column shows the people still to ask at their accounts against the requests they need until the re-measure. At a few dozen requests each, a seller's own figure shows direction only, and is for their development with their manager, never for discipline (policy items P10 and P11).

The weekly routine

About 101 minutes a week each, with the review

ActivityWhat it meansMinutes each (assumed)Times a week
R1 · FindSearch the seller's target accounts for people in the buying-group roles and add them to the seller's list, in list order35
R2 · Connect with a personal noteA connection request to a person at a target account, with a note that says why (one line about them or their work, never a pitch)3each seller's requests
R3 · EngageA thoughtful comment on a post by a person at a target account (adds something; never 'great post')46
R4 · ShareOne approved item from the content plan, with a line of the seller's own101
R5 · Move to conversationA message to an accepted connection with something useful to them, saying who it's from and offering an easy way to say no more messages; a call offered where it fits; no sequences63
R6 · Log it in the CRMEach accepted connection, conversation and meeting recorded, with the source field set to the social value26
Review · Weekly review with the managerThe seller's numbers for the week and one thing to do differently (from Inputs)101

Every request goes to a person at the seller's own target accounts, drawn in list order, with a personal note; every message is written by the seller. No tool sends, connects, comments, likes or messages for them, and nothing copies profiles or contacts in bulk: the platform's user agreement forbids it, and the policy says so. Every message says who it is from and offers an easy way to opt out, and sole traders and partnerships are left out (PECR treats a direct message as electronic mail). All twelve sellers have the platform's paid seat (from the client's quotation), because a free account allows only a handful of personal notes a month.

ModuleHoursWhat each seller producesWave 1Wave 2
M1 The profile standard1.5 hProfile rewritten to the standard and scored by both scorers02 Nov08 Feb
M2 Finding the right people at your accounts1.5 hForty people found in the buying-group roles at the seller's target accounts, in list order09 Nov15 Feb
M3 The personal note, the platform's rules and our policy1.5 hThree notes of their own; the policy signed16 Nov22 Feb
M4 Engaging and sharing1.5 hFirst comments and the first approved share23 Nov01 Mar
M5 From connection to conversation to meeting, and logging it1.5 hFirst conversations and meetings logged with the source field30 Nov08 Mar
M6 The weekly routine: owning it1 hRoutine and minutes agreed with their manager07 Dec15 Mar
Profiles

Three of twelve, before and after, against the ten-item standard

SellerBeforeAfter
Idris
Business development manager, health
Business Development Manager at Ellerholt
3 of 10 (A), 3 (B)
Business development, Ellerholt Building Services Engineers | Helping hospital estates teams plan heating, ventilation and power upgrades they can deliver around patients
10 of 10 (A), 10 (B, blind)
Joel
Regional director, Manchester
Director
1 of 10 (A), 2 (B)
Regional director, Ellerholt Manchester | Building services engineering for clients and contractors across the North West
10 of 10 (A), 10 (B, blind)
Sorcha
Associate director, mechanical (health)
Associate Director
2 of 10 (A), 2 (B)
Associate director, mechanical, Ellerholt | Ventilation and heating design for hospitals, including theatres and isolation rooms
10 of 10 (A), 10 (B, blind)

12 of 12 meet the standard; 2.6 items before, 10.0 after; scored by marketing manager and, blind, bid manager; 0 more than 1 item apart. The standard: 1. A current photo of the seller's face, taken professionally or to the same standard; 2. Headline says who they help and with what, not just a job title; 3. Current role names the business and says what it does in one line, in the customers' words; 4. About section says who they help and how, with no claim beyond the business's approved wording; 5. Experience: the current role described and the dates right; 6. One way to get in touch that works (tested); 7. Featured or pinned item points to something useful that exists, approved by marketing; 8. Background image is the business's current approved version; 9. Nothing confidential about clients; no client named without the client's permission; 10. Headline within the character limit (counted).

Target accounts

One seller each, with the buying-group roles to find

SegmentAccountsTier 1People in the roles (estimate)FoundAsked in the tallyTier 1 with 3+ connectedTwo or more roles connected
Health2045072051284 of 416 of 20
Education2045072101284 of 416 of 20
Commercial2045072091284 of 416 of 20

Accounts are described by kind here; the client's own sheet names them. The roles come from the Buying Group Map: estates or facilities director, capital projects lead, project manager and procurement.

Content

13 items over 13 weeks, every one approved before anyone shares it

WeekKindTitleWrites itApproves it
1Company articleFive questions to ask before a hospital ventilation upgradeMarketing managerTechnical director
2Someone else's article or guidancePublished guidance on heat networks, with our associate director's two-line viewAssociate director, sustainabilityTechnical director
3Case studyA university library moved from gas boilers to heat pumps, with the building open claim: evidence on file client named: permissionMarketing managerTechnical director
4Seller's own postWhat a standby generator test taught us about hospital resilienceAssociate director, electrical (health)Technical director
5Thought leadership pieceWhy we think estates teams should plan for the building they run, not the one on the drawingsManaging directorTechnical director
6Company articleElectric vehicle charging in logistics schemes: what the power supply needsMarketing managerTechnical director
7Event or webinarWebinar: planning ventilation work in occupied schoolsMarketing managerTechnical director
8Someone else's article or guidancePublished research on overheating in schools, with our view on what to check firstAssociate director, mechanical (education)Technical director
9Case studyAn office refurbishment that kept its existing risers claim: evidence on file client named: permissionMarketing managerTechnical director
10Seller's own postWater hygiene in hotels after a long closure: what to checkAssociate director, public health (commercial)Technical director
11Company articleHeat pumps in older teaching buildings: three things that decide itMarketing managerTechnical director
12Thought leadership pieceWhat we've learned from designing services for buildings that stay openManaging directorTechnical director
13Company articleLaboratory ventilation: the questions developers ask us mostMarketing managerTechnical director

The approver is never the writer. The two case studies make claims backed by evidence on file and name clients who gave permission (CAP Code 3.7 and 3.50 where it applies, a house rule everywhere). The managing director's pieces come from the Thought Leadership Programme; sellers reshare them with a line of their own.

Policy

Twelve items, agreed in time

ItemReference (checked 16 October 2026)Status
P1 Social media policy for sellingThe client's own HR policy; the solicitor where it touches the employment contractAgreed 09 Oct 2026 HR manager
P2 Who owns contacts made for work, and what happens when a seller leavesThe employment contract and confidentiality terms: for the client's solicitor (not legal advice)Agreed 09 Oct 2026 HR manager
P3 Lawful basis for recording contacts in the CRMICO, A guide to lawful basis: legitimate interests (updated 23 March 2026)Agreed 10 Jul 2026 Finance director (data protection lead)
P4 Telling contacts we hold their detailsUK GDPR Article 14(3) (legislation.gov.uk)Agreed 10 Jul 2026 Finance director (data protection lead)
P5 Emails and direct messages to contacts found on the platformICO, Electronic mail marketing (PECR): direct messages via social media are coveredAgreed 09 Oct 2026 Marketing manager
P6 No automation or bulk copying on the platformThe platform's user agreement (LinkedIn's effective 3 November 2025, section 8.2), checked against the live version on the date on InputsAgreed 09 Oct 2026 Sales director
P7 One real account per sellerThe platform's user agreement, checked against the live versionAgreed 09 Oct 2026 Sales director
P8 Claims in posts and sharesCAP Code rules 3.7, 3.47, 3.50; scope clause I h (asa.org.uk)Agreed 09 Oct 2026 Marketing manager
P9 The regulator's rules on social mediaFor example the FCA's FG24/1, Financial promotions on social media (26 March 2024); otherwise the business's own regulatorNot regulated: doesn't apply
P10 Measuring the sellers themselvesICO, Employment practices and data protection: monitoring workersAgreed 10 Jul 2026 HR manager
P11 Per-seller results never used for discipline unless the policy says soHouse rule; ICO monitoring workers guidanceAgreed 10 Jul 2026 HR manager
P12 How long the tallies and per-seller figures are keptUK GDPR storage limitation; the retention period on InputsAgreed 10 Jul 2026 Finance director (data protection lead)

Not legal advice. The five items covering the tally itself (the lawful basis and notice for contacts recorded, telling sellers what is counted, discipline and retention) were agreed before it began. The client's solicitor reviewed who owns contacts made for work when a seller leaves; their fee is the client's cost, paid directly.

How the value is shown

Measure later, sized before collecting, with Unlocks

Sized before the tally:

MeasureSizing
Conversations per request (main)Enough: 384 requests planned against 258 needed
Acceptance (process)Enough: 384 requests planned against 153 needed
MeetingsA snapshot, direction only: about 4 expected in the tally weeks against 25 needed
MeasureNowSmallest change it could show (80% power)What it can show
Main: conversations per request to people at target accounts (a reply within 28 days)2.9% (11 of 384)4.6 points (target gap 5.1)A real change of the target's size would be seen four times in five
Process: acceptance (within 14 days)27.1% (104 of 384)9.5 points (target gap 12.9)A real change of the target's size would be seen four times in five
Third: social-sourced meetings (CRM source field)4 in 8 weeksa rise of 455%Direction only: too few meetings to say more (a snapshot)
Early read: wave 1's conversations per request net of wave 23.5% and 2.2%9.2 pointsDirection only: the target gap is smaller than the smallest change it could show

3.5% wave 1 and 2.2% wave 2 now; direction only: the target gap is smaller than the smallest change it could show (smallest change 9.2 points). Tally 20 Jul to 13 Sep 2026; re-measure window 15 Mar to 9 May 2027. Different times of year, and the tally ran through the summer holidays. The re-measure isn't netted, so read it with the early read's netted change. 12 sellers have a manager who also manages the other wave: managers are asked not to pass the routine to wave 2 early; wave 2's paid seats start with its first module. Requests are counted once each has had its full 14- and 28-day windows, and meetings come from a CRM source field set up before the tally. The re-measure compares everyone with the tally and isn't netted. A change shows what the profiles, notes and routine did together, never one alone.

Unlocks: Target Account List: 12 of 12 Tier 1 accounts have 3 or more people connected; Buying Group Map: 48 of 60 accounts have someone connected in two or more roles. The routine builds both: people found and connected at each account feed the Target Account List's mapping and the Buying Group Map's roles met before a proposal.

Cost in full: £71,390 (£6,500 fee ex VAT; £39,026 sellers' routine time; £4,284 modules; £5,336 managers' reviews; £1,740 setting up; £3,596 content; £8,064 tools; £744 source field and counting; £1,200 solicitor and £900 other specialists, paid directly).

Break-even: To cover the cost in full takes about 19 extra social-sourced meetings a year, 82% more than today's 23; by the re-measure a rise under 311% could be chance, so a rise that just pays its way couldn't be told from chance. Illustration only: £87,400 a year from 23 extra meetings (capped at the owner's meetings target; 41 if the main target were reached in full), not a forecast.

How confident are we?

What this rests on

The profile standard is judged by marketing manager and, blind, by bid manager, neither a seller nor a writer of the rewrites. The routine's minutes are estimates (assumed) checked against the minutes each seller agreed; the platform's limits are assumed until checked on its help pages, and its terms were read against the live user agreement on 14 Oct 2026. The baseline is a 8-week tally of 384 requests to people at the sellers' own target accounts (the same requests as on Targets), drawn in list order (not chosen), sent the way they do now and counted once each had its 14- and 28-day windows; meetings come from a CRM source field set up before the tally, and the policies covering the tally were agreed before it began. The tally was sized before collecting. The main measure is conversations per request; acceptance is a process measure. At the early read wave 1 is compared with wave 2, which hasn't started, over the same weeks, so the season and any platform change are netted out; the re-measure compares everyone with the tally and isn't netted, and falls at a different time of year from the tally. A change shows what the routine, profiles and notes did together, never one alone. The targets, the expected rates used for sizing, the people in the roles at each account, the 50% counted as truly extra and the minutes are assumptions; the tally, meetings, conversion, gross profit, rates and costs are from the records and quotations. Per-seller figures are for development only. Policy items are not legal advice: the client's solicitor reviews contact ownership. The break-even and any £ figure are illustrations, not forecasts.

The programme is finished when it is handed over: the six modules and the weekly routine, the profile standard with each seller's scored rewrite, the 13-week content plan with its approval rules, the signed policy and the tracking sheet.

The months after

What the agent flagged from go-live to the early read, and what people decided

Fictional, continuing the example. Ellerholt chose the guided route. The agent was set up in an AI assistant the business already uses, connected read-only to the CRM (its IT confirmed this before it committed); it never touches the platform, where the sellers act and log their own activity. Every Monday it reads what each live seller logged with the social source the week before, enters it in a new dated copy of the workbook, runs the checks, and sends each seller and their own manager a short note on their own figures only. It uses the lines on Inputs: a week below 50% of the requests a seller agreed counts as missed, and 2 missed weeks in a row are flagged; any seller over the platform's limit of 100 requests a week is flagged the same week. Every four weeks it reports what the routine brings, direction only until the early read. Measuring the sellers was agreed and they were told (policy P10); the results are for development only (P11). Nothing left the business, and no decision changed, without one of its own people approving it.

DateWhat the agent flaggedWhat was decided
14 Dec
No flag
The first week after go-live: the six wave 1 sellers logged 25 requests with the social source, against the 25 they had agreed.Nothing needed. Each seller and their manager read their own note.
18 Jan
Flag
Joel logged 1 request and then none, against 3 a week, while he wrote the hospital framework bid.His manager agreed the bid came first; Joel restarted at 3 a week from 1 February. Nothing changed on the Sellers sheet.
1 Feb
Flag
Folake logged 112 requests in one week, over the platform’s limit of 100.A browser extension had been sending requests from a saved list, which the policy forbids. The sales director had it removed the same day and the marketing manager checked every seller’s browser.
22 Feb
Flag
Kwame logged no requests in the first two weeks of wave 2.His paid seat had not been assigned. The marketing manager assigned it and his manager booked a short catch-up; he logged 4 the next week.
1 Mar
No flag
The four-weekly summary: 14 conversations from 296 requests since go-live (4.7%), against 2.9% in the tally, and 3 meetings booked.Nothing needed. The sales director will judge it at the early read on 8 March, netted against wave 2, as planned.

At the 30-day check-in on 15 December, the day after the first run, Geoff and the sales director read the first notes and the log together. No rule needed changing. What the agent added. The method is unchanged: same routine, same tally, same reads. The agent made sure the routine was checked every week, seller by seller, which managers' weekly reviews were meant to do and often skipped. One flag stopped automated requests that put a seller's account and the policy at risk; two restarted a routine that had lapsed for a reason someone could fix. Conversations per request since go-live look higher than in the tally, but that is direction only: the early read, netted against wave 2, is the measure, and no £ figure is claimed.