Repair Shop Referral Programmes That Actually Work

By Sajad, Founder at Cellbot - 25 years in the tech repair industry

Published: 10 February 2026 · Editorially reviewed: 26 August 2026 · Qualified UK consumer, advertising, privacy, tax and legal review pending

Evidence basis: an original Cellbot referral-pilot register, current UK advertising, consumer and electronic-marketing guidance, and transparent attribution and incrementality formulas. It does not prescribe a reward amount or promise new customers.

A repair shop referral programme should reward a verified new-customer event under clear terms, without buying positive reviews or collecting a friend's details without a proper route. Start with a capped pilot, one eligibility rule, one share method, one reward event and a reconciliation ledger.

Measure referred jobs and compare them with an eligible non-invited baseline. Attribution shows which code was used; it does not prove the programme caused the job.

Quick answer: invite eligible customers through a lawful channel, let them share a link or code themselves, show the important terms before participation, reward only the defined qualifying event, keep review requests separate and reconcile reward cost against contribution before expanding.

!Repair shop referral measurement chain from eligible completed job through invitation, voluntary sharing, qualifying referred repair, reward and incremental decision

Download the repair-shop referral pilot register. It connects the referrer, invitation permission, code, referred customer's voluntary entry, qualifying repair, exclusions, reward liability, contribution and final decision without encouraging unnecessary personal-data collection.

Separate referrals, rewards, loyalty and reviews

These are different objects:

Affiliate publishing is different again: a publisher recommends an offer to an audience under a commercial relationship. Use the affiliate-marketing evidence playbook for disclosure, claim control and attributable commission; do not record it as a customer referral.

Never make a referral reward depend on a five-star or positive review. The Competition and Markets Authority's current unfair-commercial-practices guidance covers fake and concealed incentivised reviews. Keep referral measurement and review governance separate through the repair-shop review guide.

Define the qualifying event before choosing the reward

Write one observable event. For example:

“A first-time customer uses the valid code before payment and completes an eligible repair after the return window stated in the programme terms.”

The exact rule may differ, but it should define:

• eligible referrer;
• who counts as a new customer;
• included and excluded repairs, locations and channels;
• minimum spend only if lawful, clear and operationally justified;
• when the code must be presented;
• reward for each party;
• issue, use and expiry conditions;
• cancellations, refunds, warranty and chargeback treatment;
• self-referral, duplicate, household, staff and business-account rules;
• limit per person or period;
• suspension and fraud process; and
• programme end or change treatment.

Do not hide important terms behind the code entry or after the customer commits. The CAP Code's promotional marketing rules place responsibility for the promotion on the promoter and require fair administration and clear significant conditions.

Choose the reward from job economics

There is no universal “best” reward. Start with contribution and customer usefulness:

eligible job contribution = cash collected - parts - direct labour - payment costs - refunds - rework and warranty provision

net referred contribution = eligible job contribution - referrer reward cost - referred-customer benefit - programme delivery and fraud loss

Run base and downside cases. If the reward consumes the contribution on common repairs, narrow the eligible service, change the reward or stop. A future discount has different cash, breakage and accounting consequences from cash, a gift or immediate price reduction; obtain qualified tax and accounting advice.

Do not advertise “free repair” when exclusions, caps or required spend make that description misleading. The repair-pricing guide owns the underlying quote; the programme must not obscure it.

Use privacy-safe sharing

Avoid asking an existing customer to upload a friend's name, phone number or email address. A simpler route is to give the referrer a code or link they can choose to share. The referred person then visits voluntarily and receives the privacy and programme information directly.

If the shop sends referral invitations by email, SMS, WhatsApp or another electronic channel, treat them as marketing rather than a repair-status message. The ICO's current electronic and telephone marketing guidance explains that PECR restricts unsolicited electronic marketing and that consent, where required, must be clear, specific and based on positive action. Keep consent or other applicable evidence, channel, date and withdrawal record.

Do not insert a referral promotion into a service message and assume it remains operational. The ICO distinguishes routine customer information from messages containing significant promotional material.

The repair-shop customer-communication guide owns message purpose, consent and preference control.

Build the attribution record

Create one referral record when a code or link is issued, then connect events without duplicating customer truth:

1. eligible completed repair;
2. referrer and programme version;
3. invitation channel and permission evidence;
4. code or link issued;
5. referred person arrives voluntarily;
6. first-time-customer check;
7. quote and qualifying repair;
8. payment, cancellation or refund;
9. reward liability approved;
10. reward issued, used, expired or reversed; and
11. accounting and customer-balance reconciliation.

Use pseudonymous programme identifiers in analysis where practical. Restrict the operational record to people who need it. A public leaderboard of referrers creates a different privacy and fairness problem and is not needed for an ordinary programme.

Run a capped incremental pilot

Attribution answers “which code was recorded?” Incrementality asks “how much behaviour changed because of the programme?”

Define a representative eligible population and a dated test:

• invited cohort: receives the referral invitation under the approved rule;
• comparison cohort: otherwise eligible customers do not receive it during the test;
• common observation: same programme period plus enough time for the qualifying event and reversals; and
• exclusions: staff, duplicates, existing customers presented as new, unresolved refunds and records missing required fields.

Use a fair allocation method and avoid withholding information customers need about repairs, safety, terms or rights. The comparison concerns the promotion only.

Calculate:

qualified referral rate = qualifying referred customers ÷ eligible invited customers

comparison new-customer rate = comparable new customers ÷ eligible comparison customers

estimated incremental rate = invited rate - comparison rate

incremental programme contribution = estimated incremental qualifying jobs × net contribution per qualifying job - fixed programme cost

The estimate is only as credible as the allocation, population and data. Seasonality, other campaigns, repeat customers and code sharing beyond the cohort can confound it. State those limits.

A small pilot may be inconclusive. Report that instead of converting noise into a success story.

Control fraud without accusing customers by algorithm

Define review triggers such as:

• same customer or contact used on both sides;
• repeated device, payment, address or account evidence;
• code use above the stated limit;
• refunded or charged-back qualifying repair;
• staff-created referrals;
• bulk or automated code distribution; or
• unusual concentration requiring manual review.

A trigger should pause the reward for human review, not automatically label someone dishonest. Record the evidence, decision, communication and appeal or correction route. Apply the published terms consistently.

Do not collect more device or identity information merely because it might help fraud detection. Use the minimum evidence needed and review the privacy impact.

Reconcile reward liabilities

Track each reward state:

• pending qualification;
• approved and due;
• issued;
• partly used;
• fully used;
• expired under disclosed terms;
• reversed after an allowed event; or
• disputed and under review.

Reconcile the programme ledger with repairs, payments, refunds and accounting records. Do not report an issued credit as a completed acquisition without the underlying qualifying repair. Do not report unused rewards as a cost saving without correct accounting treatment.

Use the repair-shop accounting guide and qualified advice for discounts, vouchers, tax and liabilities.

Decide whether to expand, revise or stop

Use five gates:

Do not optimise the share rate while reward liabilities, complaints or low-quality work grow. A referral programme should amplify a trusted repair experience, not compensate for a broken one. Use the customer-experience strategy first when service recovery is the real issue.

How Cellbot fits referral operations

Cellbot can preserve customer, repair, payment and communication context according to current features and plan. Check the Cellbot pricing page. A referral code, reward liability and incremental-analysis ledger may still require a separate controlled process; verify the workflow before launch.

Cellbot cannot establish marketing permission, write lawful terms, determine tax treatment or prove incrementality. Do not use repair-status automation to send an unapproved promotion.

The author founded Cellbot and has a commercial interest.

Repair-shop referral FAQs

What reward amount works best?

There is no universal amount. Model the actual eligible repair contribution, both parties' value, programme cost, abuse exposure and tax treatment. Test a capped offer rather than copying another shop.

Should both customers receive a reward?

Either single- or dual-sided designs can be tested. The terms, commercial model and customer value should justify the choice. Do not describe one design as inherently best.

Can I ask customers to send me a friend's phone number?

Avoid that as the default. Give the customer a link or code to share so the friend chooses whether to contact the shop and sees the relevant information directly. Obtain advice for any third-party-data route.

Can the reward depend on a positive review?

No referral programme should require a positive review. Keep review invitations separate, accept honest feedback and follow current platform, CMA and advertising guidance on incentives and disclosure.

How long should the pilot run?

Long enough to observe the qualifying event, cancellations, refunds, reward issue and likely use. Set the window from the repair cycle and programme terms rather than a universal number of weeks.

How do I track referrals without dedicated software?

A small pilot can use unique codes and a controlled ledger linked to repair and payment references. It fails when duplicate files, missing permissions or unreconciled reward balances appear.

Sources, method and update note

This article was fully rebuilt on 26 August 2026. The visible byline now matches the CMS author. The review removes invented acquisition costs, fixed reward recommendations, gamification claims, universal rollout timing and unsupported automation promises. The replacement is a privacy-safe, terms-led pilot with explicit attribution and incrementality limits.

Primary references:

• CAP Code section 8: promotional marketing, Advertising Standards Authority, checked 26 August 2026
• Electronic and telephone marketing, ICO, checked 26 August 2026
• Direct marketing guidance, ICO, checked 26 August 2026
• Unfair commercial practices, Competition and Markets Authority, checked 26 August 2026
• Data-protection principles, ICO, checked 26 August 2026

Search evidence is dated 26 August 2026. An AI Overview appeared. Current auto-repair referral guides and live programmes led the standard organic results; Cellbot was absent, and the query was partially confused with the BBC programme The Repair Shop. DataForSEO SERP reference: `08261941-1339-0139-0000-41c1f829ad85`. The submitted referral phrases were omitted from the keyword-overview response, so no volume was inferred. Exa supported source and competitor discovery, not rank evidence.

Continue with the customer-experience strategy, review-request guide or customer-communication system.