What matchback analysis measures, what it misses, and how to fill the gaps

A rooster crows every morning, and every morning the sun comes up. Judged on its track record of predicting sunrises, that rooster deserves a lot of credit. Matchback analysis has a similar problem. Mail goes out, orders come in from people who got it, and the report connects the two.

That connection is real, but it can't tell you is whether the mail made those orders happen or just showed up before they did. Direct mail is one of the pricier marketing channels per piece once you add printing, postage, and list costs, so that difference shapes whether a business scales a program, trims it, or cuts a channel that's actually working. Knowing what a matchback can and can't show you keeps that call grounded in reality.

Key takeaways

• Matchback analysis connects a campaign's recipient list to later purchases, including people who never used a code or link.
• A matched order isn't the same as an added order, because some recipients would've purchased without the mail.
• Matchback can also miss real responses when customers check out with different contact details, so it isn't a reliably conservative number.
• Most matchback systems can't see purchases made at retailers or on Amazon, which means they miss a large share of what mail drives for omnichannel brands.
• QR codes and dashboards that update in real time make matchback faster, but they can't show whether the mail added revenue.
• Holdout tests show what mail added under one set of conditions, and their results can drift as offers, audiences, seasons, and other channels change.
• Pairing matchback with periodic tests and a marketing mix model gives you audience insights, point-in-time checks, and an ongoing view across every place you sell.

What is matchback analysis?

Matchback analysis connects a marketing campaign to the purchases that followed it. It compares the list of people a campaign reached against customer transaction records, then flags recipients who bought within a set period. Marketers use it most often for direct mail campaigns and catalogs, since those don't come with a click to track. The same approach shows up in digital marketing campaigns, too, where some companies match households by IP address or device ID instead of by mailing address.

How the matchback process works

Most matchback systems follow the same basic steps, whether your team runs the process in-house or uses a vendor's service.

1. Save the final mail file. Use the list of people who actually received the piece, not the original target list.
2. Import transaction data from multiple sources. Ecommerce orders, call center calls, and any point-of-sale or loyalty data all belong in the match.
3. Clean up the addresses. Many companies use USPS-certified tools here. CASS certification tests how accurately address-matching software standardizes addresses, and NCOALink updates records using change-of-address filings from people who've moved.
4. Match the records. Matchback systems connect customers using names, addresses, emails, phone numbers, or device identifiers.
5. Apply the match window. Only purchases within a set period after the mail arrives count toward the campaign.
6. Build the matchback report. Results roll up into the numbers your team reviews.

What a matchback report tells you

A typical matchback report summarizes direct mail performance in a handful of metrics.

What matchback does well

Matchback earns its place in a measurement strategy, blind spots and all, because it catches responses that other tracking misses.

• It finds buyers who skip the code. Most people who respond to mail won't scan QR codes, type in personalized URLs, or enter a promo code at checkout. Matchback can still connect their purchase to the campaign.
• It surfaces audience insights. Breaking results down by list source, geography, and customer type helps you identify which audiences to target in your next campaign.
• It gives a fast directional read. Response rates and conversion rates help you track whether campaign performance landed in the right ballpark.

Where matchback analysis falls short

The challenge is that matchback's blind spots go in both directions. It can give a campaign credit it didn't earn, and it can miss orders the campaign did drive.

It counts people who would've bought anyway

Some recipients were already planning to purchase. A loyal customer who reorders every six weeks will show up as a match if a postcard lands in week five, even though that order was coming either way. Mailing existing customers creates more of this, since they're the most likely to buy regardless.

Two terms help keep this straight:

• Attributed revenue is revenue from people who received the mail.
• Incremental revenue is revenue that wouldn't have happened without it.

A matchback gives you the first number. It can't give you the second.

It can miss real responses, too

Matching only works when the records line up. A customer who checks out with a work email, orders through a spouse's account, or recently moved may never match, even if the mail is exactly why they bought. Typos and shared addresses create more noise. So a matchback number isn't automatically conservative, and it can overcount and undercount at once.

It can't see most retail and marketplace purchases

This is the biggest challenge for omnichannel brands. Someone gets your postcard, then picks up your product at Target or—because their card's already saved there—orders it on Amazon. Unless that purchase ties back to a loyalty ID you can access, matchback systems will never find it. At Prescient, we call this a halo effect: revenue a campaign drives indirectly through another channel.

Every channel claims the same order

The person who got your mailer probably also saw a Meta ad and searched your brand name before buying. Matchback credits the mail, Meta's reporting credits Meta, and Google credits the search ad. Add up what each channel claims, and the total often beats the sales the business actually brought in. Matchback systems can't sort out those interactions across multiple channels.

Your match window shapes the answer

The window you choose decides which purchases count, so it can make the same campaign look very different.

Pick your windows before the mail goes out, and count from the in-home date rather than the send date. Choosing one after you see results makes it too easy to pick whichever flatters the campaign.

What matchback misses: an example from the Prescient platform

Those blind spots show up in real data. One omnichannel brand in the Prescient platform sells through its own site, Amazon, TikTok Shop, and national retailers. Over four weeks, it spent $65,125 on direct mail. Here's how the channel's own reporting compared to what Prescient's marketing mix model (MMM) measured.

On reported numbers alone, direct mail barely paid for itself. Our MMM measured about 6.4 times more revenue.

Where that revenue landed matters most for matchback. All of direct mail's modeled revenue showed up as halo effects, spread across every place the brand sells.

A matchback against the brand's Shopify orders could only ever see that 38% slice. Even a perfect match rate wouldn't find the other 60%, because those purchases never touched the data being matched.

Why faster tracking tools don't solve the core problem

Newer tools developed for direct mail attribution focus on speed. Assigning unique QR codes to each mailpiece, assigning IDs that sync to your CRM, and dashboards that track scans in real time all cut down on time consuming operations work. Those capabilities are genuine improvements.

They still track who responded, though, and not what the mail added. A scan you see in real time is still a scan from someone who might've bought anyway, and it still can't see the Amazon order. Here's how the common options to track mail compare.

Holdout tests help, but only for that moment

A holdout test sets aside a random slice of your mailing list and doesn't mail it. Compare purchase rates between the two groups, and the difference shows what the mail added. Marketers often run one before a big jump in mail volume.

The catch is that a holdout result captures one moment. It reflects that offer, that list, that season, and whatever else your team was running. Change any of these, and the result may not hold:

• Seasonality and broader demand trends
• The offer and creative
• The list source and how often those people have been mailed
• Spend in your other channels, which can lift or crowd out mail's impact

That's why we think of incrementality tests as locally accurate but globally inaccurate. They're worth running, but a single test shouldn't become the permanent answer for a marketing channel. If you run one, randomize the holdout before anything goes out and apply the same matching rules to both groups.

How to get a fuller picture than matchback alone

Each method in that table answers a different question, so the strongest measurement strategy uses them together. Matchback gives you audience insights, codes and tracked calls show how people respond, and holdout tests act as periodic checks on a single direct mail campaign. A marketing mix model fills in the ongoing view across every channel, which matchback can't replace. Our MMM works backward from the revenue you can observe (online, on Amazon, and in stores) and estimates what drove it, so it doesn't need a matched record to credit direct mail for an Amazon order.

Questions to ask your matchback provider

A good provider should be able to explain how its matchback systems work, not just hand you a revenue number. Ask:

• Which identifiers do you match on, and are matches exact or probabilistic?
• Do you match at the person, household, or address level?
• Do you match against the final mail file or the original target list?
• Can you separate new customers from existing ones?
• Can your service include retail, marketplace, or call center data?

Where Prescient comes in

Prescient's marketing mix model reports on direct mail alongside every other channel in your mix, so you can see what it helps generate without relying on a code, a link, or a matched record. The platform splits direct mail's revenue into base revenue and halo effects and shows where those purchases land across your site, Amazon, and retail partners. If your team trusts a holdout result, it can be incorporated as a prior, and you'll see how it affects model accuracy before you decide.

When it's time to plan, Media Forecaster shows what's likely to happen if you scale direct mail spend or shift budget into it from other channels. That turns direct mail attribution from a backward-looking report into a way to optimize your marketing strategy and your direct mail investment. Book a demo to see how Prescient measures your direct mail campaigns.

FAQs

What are the four main types of marketing analytics?

The four types are descriptive, diagnostic, predictive, and prescriptive, a framework adapted from Gartner. Descriptive analytics covers what happened, diagnostic looks at why, predictive estimates what might happen next, and prescriptive points to what you should do about it. A matchback report is mostly descriptive, since it summarizes who bought after a campaign.

What is the 7 step marketing analytics process?

There isn't one official version, but most follow a similar arc: set a business goal, choose metrics, collect data, clean it, analyze it, act on what you find, and measure the results. Matchback analysis usually fits into the collection and analysis steps for direct mail campaigns.

What is a customer targeting analysis and how does it work?

A customer targeting analysis looks at who's buying from you, and who responds to specific marketing efforts, to figure out which audiences are worth reaching next. It groups customers by traits like location, purchase history, and order value, then compares how each group performs. Matchback adds insights here by showing which segments responded to a direct mail campaign.

What is marketing data analysis?

Marketing data analysis is the work of turning raw campaign and revenue data into insights that help marketers decide what to do next. That can be as simple as comparing conversion rates across campaigns or as involved as modeling how every channel contributes to revenue over time.

How long should a matchback window be?

It depends on how long your customers usually take to buy. A low-cost impulse product might need only a couple of weeks, while a higher-consideration purchase could need 60 days or more. Many teams set a primary window plus a longer secondary one to catch delayed responses.

How do you measure the ROI of direct mail?

Start with the full costs of the campaign, including printing, postage, list rental, and any discount you offered. Then compare those costs to the revenue the mail added, not just the revenue from people who received it. Matched revenue alone can overstate success while missing purchases on other channels, so a marketing mix model that sees ecommerce, marketplaces, and retail together gives a clearer read on whether a campaign was successful.