How Do You Choose a Fundraising Agency?

A fundraising agency is one of the few vendors a charity hires that can meaningfully change its revenue line in a single year, in either direction. The selection process most boards run does not reflect that. A shortlist arrives through referral, three agencies pitch, the room picks the one that presented best, and the contract is signed before anyone has seen how the agency actually reports on a losing month.

This is the sequence we would use if we were on the buying side in 2026, including the questions that reliably separate operators from presenters.

What does a fundraising agency actually do in 2026?

The category has split. Buying an agency without naming the discipline is the first and most expensive mistake, because the four models below have different economics and rarely live under one roof at a high standard.

Paid acquisition and digital performance. Meta, Google Grants and paid search, TikTok, programmatic, landing page conversion work. Measured in cost per acquisition, return on ad spend, and second-gift rate.

Strategy and case development. Feasibility studies, case for support, campaign architecture, major gift and capital campaign counsel. Measured in pipeline built and campaign readiness.

Direct response production. Direct mail, telefundraising, face to face, monthly giving programs. Measured in net revenue per donor and attrition.

Brand and creative. Positioning, identity, message architecture, film and photo. Measured in recall and in downstream lift to the channels above.

An agency that says it does all four equally well is describing an ambition, not a capability. Decide which discipline your next twelve months actually turn on, then shortlist inside it. If the gap is positioning rather than media, start with brand before you buy traffic.

The five questions that reveal how an agency really works

Pitch decks are optimised for the room. These questions are not.

1. Show us a campaign that did not work, and the reporting you sent the client that month.

The reporting artefact matters more than the story. You are looking for whether bad news travelled early, in writing, with a diagnosis and a corrective plan. Agencies that cannot produce that document usually do not produce it for clients either.

2. Who does the work, and what percentage of their week is on our account?

Senior people pitch, junior people execute. That is not automatically wrong, but it should be priced and disclosed. Ask for named individuals, their weekly hours, and a contractual commitment that the named team does not change without your consent.

3. What is your second-gift rate across your donor acquisition clients?

First gifts are easy to buy. Retention is where fundraising economics live. The Fundraising Effectiveness ProjectSource 1: Association of Fundraising Professionals, Fundraising Effectiveness Project. Quarterly sector benchmarks on donor acquisition, retention, and revenue., run by the Association of Fundraising Professionals, publishes sector-wide donor retention and new-donor benchmarks each quarter, so both sides can talk about a real number instead of an adjective.

An agency that does not track second-gift rate is selling you volume, which is why the donor journey after the first gift belongs in the brief. You can also see how we report on mandates like this in our case studies.

4. Who owns the ad accounts, the creative files, the donor data, and the pixel?

The answer should be: you do, in writing, on day one. Agencies that hold your Meta business manager or your CRM integration hostage make switching costs the retention strategy.

5. How do you handle donor data, consent, and cross-border processing?

In Canada this means PIPEDA and, for anti-spam, CASLSource 5: CRTC, Canada's Anti-Spam Legislation. Consent rules governing electronic fundraising messages. express and implied consent rules, which carry real penalties. Registered charities also have CRA obligationsSource 4: Canada Revenue Agency, Fundraising by Registered Charities. How the CRA evaluates fundraising expenditure and cost ratios. around fundraising costs and direction and control over resources. A UK-facing program adds Fundraising RegulatorSource 6: Fundraising Regulator, Code of Fundraising Practice. The standards applying to fundraising in England, Wales, and Northern Ireland. code compliance. An agency that cannot speak to these without checking with someone should not be running your acquisition.

What should it cost, and which pricing models are safe?

Four models dominate. Only three of them are defensible.

Monthly retainer. Predictable, appropriate for ongoing strategy and channel management. Ask what happens to the fee when spend drops in a slow quarter.

Project fee. Right for feasibility studies, case development, brand work, and campaign builds with a defined endpoint.

Percentage of media spend. Common in paid acquisition, typically layered on a base retainer. The structural problem is obvious: it rewards spending more. Cap it, or pair it with an efficiency target.

Percentage of funds raised. Avoid. Both the AFP Code of Ethical StandardsSource 2: AFP, Code of Ethical Standards. Professional standards, including the prohibition on percentage-based compensation. and the CFRESource 3: CFRE International, Ethics and Accountability. The accountability standards CFRE-certified practitioners work to. accountability standards prohibit percentage-based or contingent compensation for fundraising. It is a professional ethics line, not a preference, and it also distorts which donors get pursued.

On budget, the practical question is not the agency fee in isolation but total fundraising cost ratio. The CRA guidance above sets out how it evaluates fundraising expenditure for registered charities. Model the blended ratio, including agency fees, media, and production, before you sign.

The diligence sequence that surfaces problems before the contract

Run this in order. Each step is cheap relative to the cost of the wrong agency.

1. Define the mandate in one paragraph. The discipline, the revenue outcome, the twelve-month budget, and who internally owns the relationship. If you cannot write it, you are not ready to brief.

2. Shortlist three, not seven. Two from referrals inside your subsector, one from outside it. Subsector fluency is worth more than general prestige, particularly in faith-based, health, and international development fundraising.

3. Request a written response before any meeting. Same six questions to each agency, same word limit. Written answers expose thinking that a live pitch flatters.

4. Reference-check off-list. Ask each agency for three clients, then find a fourth yourself, ideally one that left. Ask the leaver what the offboarding was like.

5. Frame the engagement as a partnership, not a purchase. The strongest agency relationships operate as one team with a shared goal: agency and charity in the same planning meetings, the same dashboard, the same definition of success.

Agree in the first 60 to 90 days on how you will work together, judged against published channel norms such as the M+R BenchmarksSource 7: M+R Benchmarks. Annual analysis of nonprofit digital fundraising performance across channels. study rather than agency averages. How an agency behaves inside a shared working rhythm tells you more than any pitch.

6. Negotiate the exit before the start. Notice period, data and asset transfer within a fixed number of days, no re-onboarding fee, and a clear statement that all accounts and creative belong to the charity.

How do you evaluate the agency once they are in?

Set the scorecard during the contract negotiation, not at the first quarterly review. For an acquisition mandate, the four numbers that matter in year one are cost per new donor, second-gift conversion rate, twelve-month donor value, and net revenue after all fees and media. For a strategy or major gift mandate, they are qualified pipeline created, proposals submitted, and average gift size movement.

Review monthly on the numbers, quarterly on the relationship. If the agency cannot explain a bad month in specific, causal terms by the second review, the problem is usually measurement infrastructure rather than effort, and that is fixable only if both sides name it early.