Product Sampling ROI: How to Measure a Sampling Campaign, With Benchmarks

Why sampling ROI is hard to measure

Product sampling sits between media and sales. A sample is an ad that the consumer uses, so its value shows up in several places: in who received it, in whether they tried it, in what they thought, and eventually in the till. Measure only one of those and the campaign looks either better or worse than it was.

The academic evidence explains why. In a study of two field experiments published in Marketing Science in 2004, Kapil Bawa and Robert Shoemaker found that free samples work through three separate effects:

• Acceleration: people who would have bought anyway start buying sooner.

• Cannibalisation: some people take a free sample instead of making a paid first purchase.

• Expansion: people who would never have considered the brand start buying it.

They also found that free samples “can produce measurable long-term effects on sales that can be observed as much as 12 months after the promotion”, and that effectiveness varied widely, even between brands in the same category. A sampling campaign judged only on the next four weeks of sales will understate its return. A campaign judged only on how many samples left the warehouse will overstate it.

The answer is to measure in layers.

The six measures that make up sampling ROI

Each layer answers one question, and each depends on the one before. A campaign that cannot prove delivery has nothing to build trial or intent figures on.

1. Delivery proof

Start with the basics: how many units went out, to which locations, on which dates. Ask your sampling partner for distribution records by site, not a single national total. The measure that matters is samples in the hands of the target audience.

2. Reach

Reach counts unique people. A sample handed to the same person twice is one reach. The useful version of this number is reach within the target audience: for a student campaign, that means first-year students in the week they arrive.

3. Trial

Trial is the share of recipients who used the product. It is the first layer that requires asking people, so the quality of the survey matters: who is asked, when, and how many reply. Ask after the product has had time to be used, and report the sample size with every result.

4. Purchase intent

Purchase intent is the most widely reported sampling KPI and the easiest to misuse. Two rules help:

• Keep the question fixed. Use the same wording and scale every time, so the result can be compared with past campaigns and with a benchmark.

• Treat intent as a leading indicator. Research by J. Scott Armstrong, Vicki Morwitz and V. Kumar found that simple intention-based estimates can be biased, “overstating or understating actual purchasing”, and recommended calibrating intentions against real sales data. An intent score tells you whether the product landed. Pair it with sales data before using it to forecast.

5. Repeat

Repeat purchase is where sampling earns its return, and it is the hardest layer to measure. Use the tools that link the sample to a later purchase: unique voucher codes, retailer or loyalty data where you have access, and a follow-up survey of the same recipients. Allow time. Bawa and Shoemaker’s 12-month finding is a reason to keep reading the data well after the campaign ends.

6. Cost per trial

Cost per trial turns the layers into a single efficiency number that can be compared with other channels:

• Cost per sample delivered = total campaign cost ÷ samples delivered

• Cost per trial = total campaign cost ÷ (samples delivered × trial rate)

• Cost per intending buyer = total campaign cost ÷ (samples delivered × trial rate × purchase intent rate)

Use the full campaign cost, including product, packaging, distribution and measurement. The further down the list you go, the closer the number gets to the cost of acquiring a customer.

Benchmarks: what good looks like

A purchase intent score means little without something to compare it with. Dig In has run 196 brand surveys for 161 brands since 2018, asking students the same core questions after they have tried a product. Across those surveys, the median brand scored:

• 85% likely to buy again in 2025 (24 surveys), and 88% in 2026.

• 96% would recommend in 2025 (20 surveys).

• 8.8 out of 10 for taste in 2025 (18 surveys).

These figures are the reference point for any student sampling result. A score well below the median is a signal to look at the product, the fit with the audience or the timing. A score above it puts the product ahead of the typical brand in a set that now covers more than 150 brands. Our benchmark page explains how the norms are built and updated.

Named campaign results

Individual results show the range. Each figure below comes from a survey of students who received the product, with the year and sample size.

NIVEA also scored 98% on “would recommend” in the same Freshers 2024 survey. The pattern across categories is consistent: when the product suits the moment, a large majority of students who try it say they will buy it again. More detail is on our results page.

Build a measurable next step into the sample

The strongest sampling campaigns give the student something to do next, and make that action trackable. A registration page, a QR code or a voucher turns a sample into a measurable response.

Santander UK’s 2025 campaign shows the effect. Its mailer recorded 1,053 registrations from 4,037 views. The same partnership’s 500,000 welcome totes were still being carried by 96% of surveyed students weeks later (n=50,000), according to Santander’s End of Campaign Report 2025.

A measurement checklist

Before the campaign:

1. Agree the primary KPI (trial, intent, repeat or registrations) and the benchmark you will judge it against.

2. Fix the survey questions and scales in advance.

3. Build in a tracked next step: a code, a link or a registration.

4. Check compliance. Section 8 of the CAP Code requires promoters to make sure product samples are safe, and alcohol must not be available on promotion to anyone under 18. Age-restricted products need age-checked distribution.

After the campaign:

1. Report delivery proof by site and date.

2. Report survey results with sample sizes, against the benchmark.

3. Calculate cost per sample, cost per trial and cost per intending buyer.

4. Keep tracking codes and sales for up to 12 months.

Dig In’s approach to each of these steps is set out in how we measure.

What this means for brands

Set the KPI before the first sample ships. Many arguments about sampling ROI are really arguments about what the campaign was meant to do.

Judge intent against a benchmark, and sales over time. Intent tells you quickly whether the product landed with the audience. Repeat purchase tells you whether it paid back, and that can take months to show.

Pay for measurement as well as distribution. A partner that can prove delivery, survey recipients at scale and compare the result with hundreds of earlier campaigns gives you a number you can defend in a budget meeting. Dig In has collected over a million survey responses since 2016 for that reason.

Questions brands ask

How do you measure product sampling ROI?

Measure six layers in order: delivery proof, reach, trial, purchase intent, repeat purchase and cost per trial. Compare survey results with a benchmark, and track sales through codes or retailer data for up to 12 months, since sampling effects can last that long.

What KPIs should a sampling campaign track?

Samples delivered by location and date, unique people reached in the target audience, trial rate, purchase intent, repeat purchase and cost per trial. Add a tracked action, such as registrations or code redemptions, where the campaign allows it.

What is a good purchase intent score after product sampling?

For student sampling, the Dig In median brand scored 85% “likely to buy again” in 2025 across 24 surveys, and 88% in 2026. Scores well below that suggest a problem with product fit or timing.

How do you calculate cost per trial?

Divide the total campaign cost by the number of people who tried the product. That is samples delivered multiplied by the trial rate measured in a recipient survey.

How long does product sampling affect sales?

Longer than most campaigns are measured for. A 2004 study in Marketing Science by Bawa and Shoemaker found measurable effects on sales as much as 12 months after a free sample promotion.

To plan a sampling campaign with measurement built in from day one, talk to the Dig In team.

Sources

• Kapil Bawa and Robert Shoemaker, INFORMS, The Effects of Free Sample Promotions on Incremental Brand Sales (Marketing Science 23(3), 345 to 363) (2004).

• J. Scott Armstrong, Vicki G. Morwitz and V. Kumar, The Wharton School, Sales Forecasts for Existing Consumer Products and Services: Do Purchase Intentions Contribute to Accuracy? (working paper) (2000).

• Committee of Advertising Practice / ASA, CAP Code, Section 08: Promotional marketing (2026).