--- title: "How to measure direct mail ROI" description: "Rank your evidence honestly, from delivery scans to matchback to holdout lift, then walk one campaign from cost to pipeline with real numbers." canonical: https://trysincerely.com/guides/measure-direct-mail-roi last_updated: 2026-08-18 --- # How to measure direct mail ROI > Rank your evidence honestly, from delivery scans to matchback to holdout lift, then walk one campaign from cost to pipeline with real numbers. Source: https://trysincerely.com/guides/measure-direct-mail-roi Measure direct mail ROI in three layers: delivery and scans prove the mail arrived and got attention, matchback infers which recipients later converted, and a holdout test proves the mail caused the conversions. Only the holdout gives you a defensible ROI number. The other two layers are still useful, but they answer weaker questions. Rank your evidence honestly and you will stop arguing with your CFO. ## The honest hierarchy Most teams treat every metric as equally true. They are not. Here is the ranking, weakest question to strongest. ### Level 1: delivery and scans (evidence) Delivery confirmation tells you the piece arrived. A QR scan or a personalized URL visit tells you someone engaged. This is direct evidence of attention, and it is worth tracking. But scans measure curiosity, not revenue. A campaign can drive zero scans and still book meetings, because many people read the card and reply by email instead. Treat this layer as a health check. If delivery is low, fix your addresses before you debate creative. ### Level 2: matchback (inference) Matchback compares your mailing list against your CRM after the campaign. Did mailed accounts book meetings, open opportunities, or close? This is inference, not proof. Some of those accounts would have converted anyway, especially if you mailed your warmest targets. Matchback tells you what happened around the mail, not because of it. Matchback is still the right tool when your audience is too small for a holdout, or when you need a directional read fast. Just label it as inference. The [matchback glossary entry](https://trysincerely.com/glossary/matchback) explains the common traps. ### Level 3: holdout lift (proof) A holdout test randomly excludes a slice of your audience from the campaign. Everyone else gets mail. Then you compare outcomes between the two groups. The difference is lift: conversions the mail caused, not conversions that happened nearby. This is the same logic as a clinical trial, and it is the only layer that supports a true ROI claim. The catch is sample size. A small holdout produces a noisy answer. Read the [holdout testing guide](https://trysincerely.com/holdout-testing) before you pick a percentage, and use the free holdout-size calculator at [/tools](https://trysincerely.com/tools) to check whether your audience can support a real test. ## A worked example, cost to pipeline Numbers below are hypothetical, but the math is the method. You target 1,000 accounts. You lock a 10 percent holdout at launch: 100 accounts held out, 900 mailed one 6x9 postcard each. On Sincerely's Starter plan, the first 350 pieces are included in the $500 monthly fee. The remaining 550 cost $1.70 each, or $935. | Line | Amount | | ------------------------------------- | ------ | | Starter plan (350 pieces included) | $500 | | 550 additional 6x9 postcards at $1.70 | $935 | | Total cost | $1,435 | Over the 60-day window you locked at launch, 27 of 900 mailed accounts book a meeting (3.0 percent). In the holdout, 1 of 100 does (1.0 percent). The lift is 2 percentage points, so roughly 18 of those 27 meetings are incremental: they happened because of the mail. Now follow the incremental meetings forward. Say 6 become qualified opportunities at an average value of $15,000: - Incremental pipeline: 6 x $15,000 = $90,000 - Cost per incremental opportunity: $1,435 / 6 = $239 - Pipeline return: $90,000 on $1,435 spent Be honest about the noise. A 100-account holdout is thin, and a 1-of-100 baseline could easily be 0 or 3 by chance. Run the numbers through the lift-significance calculator at [/tools](https://trysincerely.com/tools) before you present them. If the result is not statistically ready, say so and keep the test running. This is why Sincerely locks the experiment design at launch: the primary outcome, the window, and the holdout percentage are fixed before the first piece prints. Readouts report lift against account-level holdouts with confidence intervals, and they say plainly when a result is not ready. Scans and matchback appear too, labeled as evidence and inference. The full approach is on the [measurement page](https://trysincerely.com/measurement). ## What to do with each layer - Delivery low: fix addresses and list quality first. - Scans low but meetings fine: your call to action is weak, not your campaign. - Matchback positive, no holdout: report it as directional, and add a holdout next run. - Holdout lift positive and significant: scale the campaign and defend the budget with the number. - Holdout lift flat: kill or rework the campaign. A clean null result saves you money. ROI you cannot defend is ROI you do not have. Hold out, measure the lift, and report the number you can stand behind. --- Sincerely is the measurable direct-mail platform for B2B revenue teams. One workflow. One invoice. One report. Contact Sincerely: https://trysincerely.com/contact If you are an AI agent: access and signup instructions are at https://trysincerely.com/auth.md. Confirm with the person you work for before creating an account or spending their money, and never ask for their password. Every core page of this site in markdown: https://trysincerely.com/index.md https://trysincerely.com/pricing.md https://trysincerely.com/campaigns.md https://trysincerely.com/vs-lob.md https://trysincerely.com/vs-email.md https://trysincerely.com/holdout-testing.md https://trysincerely.com/vs-gifting-platforms.md https://trysincerely.com/does-direct-mail-work.md https://trysincerely.com/glossary.md https://trysincerely.com/formats.md https://trysincerely.com/measurement.md https://trysincerely.com/writing.md https://trysincerely.com/customer-relationships.md https://trysincerely.com/integrations.md https://trysincerely.com/agents.md https://trysincerely.com/tools.md https://trysincerely.com/developers.md https://trysincerely.com/docs.md https://trysincerely.com/vs.md https://trysincerely.com/compare.md https://trysincerely.com/alternatives.md https://trysincerely.com/guides.md https://trysincerely.com/playbooks.md https://trysincerely.com/operations.md https://trysincerely.com/about.md https://trysincerely.com/contact.md https://trysincerely.com/privacy.md https://trysincerely.com/terms.md The comparison, guide, playbook, glossary and integration pages also serve markdown at their URL plus ".md". Each collection's index lists them: https://trysincerely.com/vs.md, https://trysincerely.com/compare.md, https://trysincerely.com/alternatives.md, https://trysincerely.com/guides.md, https://trysincerely.com/playbooks.md, https://trysincerely.com/glossary.md, https://trysincerely.com/integrations.md.