Turn Customers into Proof and Advocacy

Task

Create customer advocacy and reference program.

Summary

Build a disciplined system for references, reviews, stories, community participation, and customer evidence.

Build a Customer Advocacy and Reference Program Buyers Can Trust

Task ID: S5-06

A customer advocacy program turns verified customer results into credible proof for buyers, partners, and users. The work is not complete when a company has collected praise. It is complete when relevant, permissioned customers and evidence can be matched to real buying questions without misleading prospects, exhausting advocates, or creating a manual scramble.

The problem is not a lack of praise

A growing software company often has satisfied customers but no dependable way to use their experience.

A partner asks for a customer in the same industry. An enterprise prospect wants to speak with someone who completed a similar implementation. A marketplace buyer looks for recent independent reviews. Sales searches old email threads for an approved logo. Marketing asks the same two customers for another quote, webinar, or reference call.

The company appears to have customer proof, but the proof is not organized, current, permissioned, or broad enough to support its sales channels.

This becomes more serious as the company adds partners, marketplaces, product-led growth, expansion sales, and enterprise deals. More people need customer evidence, often without the founder or original account owner present. A few informal relationships cannot support that demand.

The operating principle is straightforward:

Build customer advocacy as a managed system for permission, evidence, matching, and mutual value—not as a campaign to collect compliments.

A credible program should be able to answer four questions quickly:

1. Which customers have achieved a result that can be verified?
2. What have they agreed to do or allow?
3. Which buyer, market, product, or use case can their experience credibly support?
4. How often can the company involve them without damaging the relationship?

This work belongs after the company can deliver a dependable customer result. Advocacy cannot compensate for weak onboarding, poor adoption, unresolved service failures, or a product that does not produce the promised outcome. It makes evidence of value easier to find and use; it does not manufacture the value itself.

What the evidence supports

Research supports the importance of referrals and customer reviews, but it also shows why simple claims such as “advocacy always drives growth” are too broad.

A study of roughly 10,000 customers at a German bank found that customers acquired through a referral program had higher retention and were, on average, at least 16% more valuable than comparable non-referred customers. The difference varied across customer segments, and the authors explicitly recommended selective rather than uniform referral strategies. The result is useful evidence that referrals can improve customer quality, but it is not a universal software-industry benchmark.

Later research on the same general setting found evidence for two mechanisms: referrers can produce a better match between customer and company, and the social relationship between referrer and referred customer can reinforce the new relationship. Experienced referrers were also associated with higher-margin referrals. This suggests that the value of advocacy may come from fit and informed introductions, not merely from lower acquisition cost.

Online reviews also affect buying behaviour, although the strength of the effect depends on the product, platform, and review measure. A meta-analysis covering 1,532 effect sizes from 96 studies found a positive average relationship between electronic word of mouth and sales. Review volume was generally more strongly associated with sales than average sentiment, while high disagreement among reviews could weaken the effect. Another meta-analysis found that the effect varied with factors including product trialability, market competition, and the independence and trustworthiness of the review platform.

Earlier research comparing book sales and reviews on Amazon and Barnes & Noble found that improved reviews were associated with higher relative sales and that one-star reviews had a greater effect than five-star reviews. The setting was consumer books rather than business software, but the finding illustrates an important principle: buyers do not treat all customer evidence equally, and negative evidence can carry disproportionate weight.

The evidence is much thinner for the isolated effect of formal business-to-business reference calls, case studies, executive advisory boards, and branded advocate communities. These practices are common, but public causal studies separating their effect from product quality, customer health, sales skill, account size, and brand strength are limited. A company should therefore measure its own reference program instead of importing broad revenue claims from vendors or case studies.

Nor should willingness to recommend be treated as sufficient proof that a customer will advocate or that the company will grow. A longitudinal study of 21 firms and more than 15,500 interviews failed to reproduce claims that Net Promoter Score was clearly superior to other customer measures as a predictor of revenue growth. More recent longitudinal work across airlines, supermarkets, and insurers reached a similar conclusion.

A satisfaction score can help identify candidates. It cannot replace evidence that the customer achieved a result, fits the buyer situation, has permission to participate, and is actually willing to do so.

There is also a hard legal and credibility boundary. In the United States, the Federal Trade Commission’s Consumer Reviews and Testimonials Rule took effect on October 21, 2024. It prohibits practices including fake reviews, sentiment-conditioned incentives, certain undisclosed insider reviews, deceptive company-controlled review sites, review suppression, and the commercial purchase or sale of fake social indicators. Incentives for reviews are not automatically prohibited, but they cannot be conditioned on positive or negative sentiment, and material connections may need to be disclosed.

The conclusion from the research is not “collect as much praise as possible.” It is:

Build a varied, honest, relevant body of customer evidence, and measure whether it helps the right buyers make better decisions.

Design the program around different proof jobs

Reference customers, review sites, case studies, and community advocates are related, but they do different work. Treating them as interchangeable creates bad requests and misleading measures.

The company should start by defining the proof jobs it needs to perform.

For a marketplace, recent independent reviews may matter most. A channel partner may need customer stories that explain implementation, integration, and partner involvement. An enterprise seller may need a reference who can discuss security review, rollout, adoption, and support. A product-led motion may benefit more from an active peer community than from scheduled reference calls.

A useful definition is:

A referenceable customer is a current customer with a relevant and reasonably verified experience who has explicitly permitted at least one defined form of external participation and remains available within agreed limits.

Under that definition, a customer name in a presentation is not automatically a referenceable account. Neither is a high satisfaction score, an expired case study approval, an employee’s personal relationship, or a customer who agreed to one private call two years ago.

The working target of 10 or more referenceable accounts can be a sensible starting point for a pilot, but research does not establish it as a universal benchmark. The right number depends on how many buyer situations the program must cover.

A company selling one product to one narrow market may gain more from a small, highly relevant set than from a larger undifferentiated list. A company serving several industries, countries, use cases, implementation models, and customer sizes will need broader coverage. A regulated enterprise buyer may require a reference with similar governance and procurement conditions, while a smaller self-serve buyer may rely primarily on reviews and community discussion.

The target should therefore be tested against a coverage matrix:

Ten accounts that all represent the same industry, use case, and enthusiastic executive may provide less useful coverage than a smaller but more varied portfolio.

Build an operating system, not a list of names

A dependable program moves customers from verified outcomes to controlled, reusable forms of participation.

flowchart LR
    A[Customer reaches a verified outcome] --> B[Screen for fit, health, and risk]
    B --> C{Customer is suitable}
    C -->|No| D[Continue customer work]
    C -->|Yes| E[Discuss mutual value and options]
    E --> F[Record written permission and limits]
    F --> G[Create review, story, reference, or community role]
    G --> H[Match proof to buyer needs]
    H --> I[Track use, results, and advocate load]
    I --> J[Refresh, expand, pause, or retire]

Text description: a customer enters the program only after producing a credible result. The company then checks suitability, agrees on a form of participation, records permission, uses the evidence selectively, monitors the burden, and regularly refreshes or retires it.

Start with one business purpose

The first program should solve a defined problem. Examples include improving reference coverage for a priority industry, gathering credible reviews for a marketplace launch, supporting partner-led deals, or producing evidence for a new enterprise use case.

A broad goal such as “increase advocacy” does not tell the team whom to recruit, which proof to create, or how to judge success.

Choose a purpose that connects to an observable buying or customer problem. For example:

Provide relevant customer evidence for enterprise prospects evaluating implementation risk in the company’s priority market.

That purpose implies a need for implementation stories, suitable reference contacts, approved outcome evidence, and a way to track which opportunities used them.

Screen candidates with more than satisfaction

The customer-success, sales, marketing, support, and product teams should identify candidates together. A practical screen considers:

• evidence of a completed or progressing customer outcome;
• product adoption and appropriate use;
• renewal or relationship health;
• unresolved service, security, billing, or contractual issues;
• fit with an important customer segment or buying situation;
• the strength and availability of an individual spokesperson;
• the customer’s communications and confidentiality policies;
• previous advocacy activity and current request load.

This is a decision aid, not an automatic score. A high-value customer with an unresolved escalation may be inappropriate today. A smaller customer with a clear, well-documented result may be more useful than a famous logo with little substance.

Net Promoter Score, customer satisfaction, usage, renewal history, and customer lifetime value can support the screen, but none should act as the sole qualification rule. The research on Net Promoter Score is a warning against using one survey response as a substitute for business evidence.

Ask after value is visible

The best time to discuss advocacy is usually after the customer can describe a meaningful result, not simply after signing a contract or completing onboarding.

The account team should first confirm what changed:

• What problem was the customer trying to solve?
• What did the customer implement or change?
• What result can be observed?
• Over what period?
• Which other factors contributed?
• What remains incomplete or difficult?

These questions make the eventual story more credible and reduce the risk of publishing claims that cannot be supported.

A recent working paper studying restaurant review monitoring found that a system linking negative reviews to operational work tickets was associated with improved ratings and more positive review content, particularly where reviews had identified weak areas. The study is specific to restaurants and should not be generalized mechanically to software, but it supports the principle that feedback becomes more valuable when it triggers operational correction rather than reputation management alone.

Offer several ways to participate

Do not ask every advocate for a case study, public quote, event appearance, review, referral, and reference call.

Offer a ladder of participation:

• private product feedback;
• advisory group participation;
• an independent review;
• an approved anonymous quote;
• name or logo use;
• a written or video case study;
• a private prospect call;
• a webinar or event appearance;
• peer support or community leadership.

The customer should be able to choose a level that fits their authority, interest, and employer policy.

Reciprocal value does not have to mean cash. It may include professional recognition, education, access to product specialists, early product information, peer networking, opportunities to develop speaking skills, or a stronger public profile for the customer’s work. Atlassian’s Champions program, launched in its current form in April 2025, separates community contribution from internal customer success leadership and offers recognition, support, and development opportunities. The lesson is not to copy its program; it is to give advocates clear roles and reasons to participate beyond helping the vendor sell.

Benefits must not be used to buy prescribed sentiment. Under the FTC rule, a company cannot condition an incentive for a consumer review on the review being positive or negative. Unexpected material relationships may also require clear disclosure.

Record permission precisely

Permission should state what the customer has approved, not merely that the customer is “a reference.”

At minimum, record:

GitLab’s public customer-advocacy handbook illustrates several useful operating practices. It distinguishes case studies, videos, speaking opportunities, earnings references, customer names and logos, advisory programs, and other forms of participation. It says references should be secured in writing, maintains centralized evidence and approval records, and requires advance planning for customer speakers. These are examples of operational discipline, not universal staffing or production benchmarks.

Written consent matters because people change jobs, products change, companies merge, results become outdated, and customer communications policies evolve. The record should be reviewed on a defined cadence and whenever there is a material account change.

Create a searchable source of truth

The central record should connect:

• the customer account;
• approved contacts;
• customer segment and industry;
• product and use case;
• implementation method;
• verified outcomes;
• available proof assets;
• permitted activities;
• request history;
• current customer health;
• owner;
• expiry or review date.

Sales and partners need a simple request process. The program owner needs authority to reject a poor match, protect a heavily used advocate, or pause an account.

The matching process should consider relevance before logo prestige. A well-matched customer should be able to discuss the buyer’s real concern. An impressive but unrelated customer may create recognition without reducing uncertainty.

Keep review campaigns independent and neutral

Review sites are valuable partly because the vendor does not control each response. The program should preserve that independence.

A defensible process invites customers who have genuine, relevant experience; uses neutral language; allows positive, mixed, or negative feedback; follows the platform’s eligibility rules; and records any incentive or disclosure requirement.

G2 says its reviews undergo automated and manual moderation, identifies incentivized reviews, and states that eligibility for an incentive cannot depend on whether the opinion is positive or negative. Gartner Peer Insights requires reviewers to have relevant professional experience and pass identity and role checks. Its vendor guidance also tells companies to invite people with direct experience, offer incentives consistently rather than only to happy customers, and avoid interfering with reviewer validation.

Companies should not ask only customers expected to provide five-star reviews, pre-write the review, submit it on the customer’s behalf, pressure the reviewer to change criticism, or hide negative feedback.

The legal and commercial risk is not theoretical. The FTC alleged that Fashion Nova suppressed reviews below four stars while representing its displayed reviews as customer opinion. The company agreed to a $4.2 million settlement, and the FTC later distributed nearly $2.4 million to affected consumers. Although the case concerned consumer retail, it demonstrates the wider principle: a review program becomes deceptive when it presents a selected picture as the full customer record.

Build case studies from evidence

A strong case study should include:

• the customer and relevant context;
• the original problem;
• the previous process or baseline where available;
• what the customer changed;
• the implementation scope and timeframe;
• the customer’s measurable and qualitative results;
• important limits, dependencies, or continuing work;
• quotations approved by the named speaker;
• the date the facts were verified;
• customer approval for each intended use.

The company should distinguish measured outcomes from estimates and interpretations. “Reduced processing time from four days to one day during the first quarter after rollout” is stronger than “transformed productivity.” When a result is an internal estimate, say so. When several changes contributed, do not attribute the entire result to the product.

Case studies also need retirement rules. A story based on a discontinued product, outdated pricing model, departed executive, or materially changed customer relationship may no longer be safe or useful.

Protect community advocates from becoming sales inventory

Community advocates may answer questions, host events, mentor users, share workflows, or provide product feedback. Their peer credibility depends on having room to be honest and useful.

Do not count every active community member as a sales reference. Some may be comfortable teaching users but unwilling or unauthorized to discuss procurement, business outcomes, security, or commercial terms. Others may be consultants, employees of partners, or independent experts rather than end customers.

Maintain separate permissions and measures for community contribution and customer-reference activity. A healthy community should create value for participants even when it does not produce an immediate opportunity.

Learn from public operating examples

GitLab’s public handbook shows what happens when customer advocacy becomes an operating function rather than an informal marketing task. Its program covers customer stories, approved logos and names, earnings references, speakers, advisory groups, reviews, and other formats. It uses selection criteria tied to sales priorities, account characteristics, use cases, and customer outcomes. It also maintains centralized records and written approvals.

The useful lesson is not GitLab’s exact organization or content volume. It is the separation of several decisions that smaller companies often collapse into one:

• Is the customer suitable?
• What story or evidence is strategically useful?
• What has the customer approved?
• Where may it be used?
• How much lead time and customer effort are required?
• Who owns the final decision?

Atlassian’s community model teaches a different lesson. Its Champions program provides distinct paths for people who contribute to the external community and those who help users inside their own organizations. It frames participation around connection, mentoring, learning, and better ways of working—not only vendor promotion.

The Fashion Nova enforcement action provides the counterexample. A program can produce an attractive rating while destroying the conditions that make ratings valuable. Suppressing inconvenient evidence may improve the displayed number temporarily, but it weakens buyer information, creates regulatory exposure, and makes every positive review less credible.

Together, the examples show three parts of good work: operating discipline, reciprocal participant value, and evidence integrity.

Measure evidence, coverage, and use

The primary measure—referenceable customers—is useful only when its definition is strict.

Count an account when:

• a current customer relationship exists;
• the relevant outcome or experience has been checked;
• at least one external activity is explicitly approved;
• the approved contact or asset is current;
• the account is not paused by a material issue;
• the record identifies the situations for which it is relevant.

Do not count a customer separately for every quote, review, contact, or case study. Report both unique referenceable accounts and available proof assets.

The working target of 10 or more referenceable accounts should be treated as an initial operating target. It can demonstrate that the company has moved beyond one-off relationships, but it does not prove sufficient coverage. Report the count alongside the dimensions that matter.

Revenue influenced by references can be reported, but “influenced” should not be presented as “caused.” Larger, more promising opportunities are often more likely to receive references, and the strongest customers are more likely to become advocates. A simple comparison between deals with and without references can therefore exaggerate the program’s effect.

A better analysis records the opportunity stage, buyer concern, account type, proof used, and outcome. Over time, the company can compare similar opportunities and examine whether appropriate references are associated with shorter decision time, higher stage conversion, fewer stalled evaluations, or improved partner effectiveness. Even then, the conclusion should remain cautious unless the design rules out other explanations.

The program should also measure failure demand: requests it cannot fulfil. Repeated requests for references in a specific industry, region, integration, or customer size reveal where the sales strategy has moved beyond the available customer evidence.

That is valuable information. A gap may mean the company needs more advocates, but it may also mean it is pursuing a market where it has not yet produced enough successful customer outcomes.

Avoid false completion and decide when the program is ready

Customer advocacy work often looks complete before it can support growth.

Common signs of false completion include:

• a spreadsheet of customer names without documented permission;
• ten accounts that all represent the same buyer situation;
• a library of stories with no verified outcomes;
• a high average review score created through selective invitations;
• one or two customers handling most reference calls;
• community members counted as references without their agreement;
• logos still in use after approval has expired;
• customer quotes detached from their original context;
• sales teams contacting advocates without coordination;
• influenced revenue reported as if advocacy caused the sale;
• negative feedback collected but never routed to product, support, or delivery owners.

The major tradeoffs require judgment.

Control versus credibility. A company controls a case study’s structure but not an independent review or reference conversation. More control can make evidence clearer, but too much control makes it less persuasive.

Volume versus relevance. A large review count or advocate list can show activity. A buyer still needs evidence from a comparable situation.

Speed versus governance. Sales wants immediate access to customers. Customers need notice, context, preparation, and protection from repetitive requests.

Incentives versus authenticity. Recognition and reciprocal benefits can support participation. Benefits tied to prescribed sentiment undermine trust and may violate law or platform rules.

Public proof versus confidentiality. Some of the most valuable enterprise customers cannot permit public logos or metrics. Anonymous stories and tightly controlled private references may still be useful, but their limits must be clear.

Central ownership versus local relationships. A central owner protects consistency and advocate load. Account owners retain essential knowledge of customer health and internal politics. The program needs both: central rules and records, with account-team input before each significant request.

The company is ready to depend on the program when the following conditions are true:

• priority buying situations have relevant customer proof;
• every active reference has current, recorded permission;
• outcomes and claims are reasonably verified;
• sales and partners use a defined request process;
• customer load and contact frequency are controlled;
• review sourcing is neutral and compliant;
• case studies and permissions have refresh dates;
• negative feedback enters an operating improvement process;
• the company can distinguish program activity from business impact;
• no single customer relationship is carrying the program.

At that point, the result is more than 10 names. It is a customer-evidence system that can support additional sales channels without sending every request back to the founder, the original salesperson, or the same exhausted advocate.

The final decision is not whether the company has enough praise. It is whether the company can provide the right proof, to the right buyer, with the customer’s informed permission, at a sustainable level of effort. If it cannot, the program should remain a focused pilot while the company closes its outcome, coverage, permission, or operating gaps.

Sources

Primary and official sources

• Federal Trade Commission, “The Consumer Reviews and Testimonials Rule: Questions and Answers.”
• Federal Trade Commission, “Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials.”
• Federal Trade Commission, Fashion Nova enforcement and refund materials.
• GitLab, “Customer Advocacy at GitLab.”
• Atlassian, “The Atlassian Champions Program Has Launched” and Atlassian Community program information.
• G2, “Community Guidelines.”
• Gartner Peer Insights, vendor and review-moderation guidance.

Open research

• Schmitt, Skiera, and Van den Bulte, “Referral Programs and Customer Value,” Journal of Marketing, 2011.
• Van den Bulte, Bayer, Skiera, and Schmitt, “How Customer Referral Programs Turn Social Capital into Economic Capital,” Journal of Marketing Research, 2018.
• Babić Rosario, Sotgiu, De Valck, and Bijmolt, “The Effect of Electronic Word of Mouth on Sales,” Journal of Marketing Research, 2016.
• You, Vadakkepatt, and Joshi, “A Meta-Analysis of Electronic Word-of-Mouth Elasticity,” Journal of Marketing, 2015.
• Chevalier and Mayzlin, “The Effect of Word of Mouth on Sales: Online Book Reviews,” National Bureau of Economic Research working paper and subsequent journal article.
• Keiningham, Cooil, Andreassen, and Aksoy, “A Longitudinal Examination of Net Promoter and Firm Revenue Growth,” Journal of Marketing, 2007.
• Dawes, “Net Promoter and Revenue Growth: An Examination Across Three Industries,” 2024.
• Cao, He, and Jin, “From Complaint to Action: Technology-Enabled Quality Improvement from Consumer Reviews,” National Bureau of Economic Research working paper, revised July 2026.