Check packaging marketing measurement before you report results

What to take away

Packaging marketing measurement rarely settles whether packaging works. The decision in front of you is which format to keep, which to retire and which supplier to renew. That choice needs evidence you can defend to a finance director who has never watched an unboxing video.

This guide covers the metrics worth tracking, the attribution methods that survive scrutiny and the reporting rhythm that keeps the numbers current. It is written for ecommerce and brand teams in England, though most of the measurement logic travels.

Choose the metrics before the format

Packaging sits across two budgets: marketing and operations. That split is why measurement arguments start badly. Marketing wants recall and sentiment. Operations wants cost per unit and damage rates. Both are legitimate, and one report has to hold them apart.

Start by naming the decision the data will inform. If you are choosing between two mailer formats, cost per delivered unit and damage rate decide it. If you are testing a printed insert, the question is whether it lifts repeat purchase within 90 days. Those are different studies with different sample sizes.

Write the decision down before you brief a supplier. It forces the team to agree what "working" means, and it stops the post-campaign scramble to find a metric that happens to look good.

Six metrics tied to a format decision

A practical set for most direct-to-consumer operations covers six areas. Cost per delivered unit includes the pack, the filler and the outer box. Damage and return rate covers items that arrived unusable. Repeat purchase rate at 90 and 180 days is the clearest brand signal a box can influence.

Unboxing share is the proportion of orders where a customer posts or tags the packaging. Insert redemption covers codes and QR scans. Sentiment comes from review text, read by a human, not a score alone.

If you are still choosing which of these to put on the front page, our guide to packaging marketing key metrics explains which measures suit subscription, gifting and low-margin retail models.

Separate leading from lagging measures. Damage rate and insert scans appear within weeks and tell you whether the pack is performing. Repeat purchase takes a quarter or more and tells you whether it mattered. Reporting both in one table without labelling which is which invites impatience from whoever reads the numbers first.

Sample size is the practical constraint. If a brand ships 300 orders a month, a 90-day repeat measure rests on 900 orders, and a small gap between packs will not clear the noise.

A team splitting 900 orders evenly between two formats is comparing 450 against 450, which is workable for a large effect and unreliable for a fine one.

Two cautions. First, unboxing share is a vanity trap if the sample is tiny; a handful of posts from a launch week proves nothing. Second, sentiment scores drift with review platform changes, so read the comments. A rising score with angrier wording is a warning, not a win.

Cost, compliance and the numbers finance asks for

Before any format reaches a customer, it has to clear the legal baseline. The Packaging (Essential Requirements) Regulations 2015 set limits on heavy metals and require packaging to be minimised by weight and volume, which is checkable against your specification sheets at the 2015 essential requirements regulations.

Cost modelling then needs a tax line. Plastic Packaging Tax applies to plastic packaging manufactured in or imported into the UK with less than 30% recycled content. The GOV.UK guidance sets out who pays and how to account for it at Plastic Packaging Tax guidance.

For a mid-sized brand, that can shift the pack cost enough to change a format decision. A team paying £400 a month for a plastic mailer may find a recycled-content version cheaper overall once the tax line is removed.

Cost reporting needs the same discipline as the metrics. Split packaging spend into pack materials, fulfilment labour, tax and measurement, so a format change can be compared on total cost rather than pack price alone. A cheaper mailer that raises damage claims is not cheaper.

Attribute the unboxing without fooling yourself

Attribution is where packaging measurement earns or loses credibility. No method will isolate the box the way a paid search click can be isolated, because the box arrives alongside the product, the courier and the delivery email.

State the method once and stay consistent. If you change method mid-year, the trend line breaks and every comparison becomes a guess.

The four methods worth using

Self-reported attribution asks customers at checkout or in a follow-up survey how they heard about you. It is cheap, it captures word of mouth, and it is biased towards whatever happened most recently. Use it as a signal, not a total.

Holdout testing is the strongest option available. You ship a plain pack to a randomly selected slice of orders for a set period and compare repeat rates against the branded pack. It is uncomfortable with customers and it is the only design that supports a causal claim.

Matched-market or cohort comparison works when you cannot run a holdout, for instance across regions. It is weaker because the groups differ in ways you did not choose.

Last-touch and multi-touch models borrowed from digital analytics will undercount packaging, because the box rarely appears in a click path. If you use them, label the packaging contribution as modelled, not measured.

Our overview of packaging marketing attribution methods walks through sample sizes and test lengths for each approach. Whichever you pick, publish the confidence interval or the sample size beside the result. A repeat-rate lift of two percentage points from 40 orders is noise.

The table below gives typical figures; treat them as a starting point.

A worked example (illustrative figures for a mid-sized direct-to-consumer brand) shows why. A brand ships 6,000 orders a month and splits them 3,000 branded against 3,000 plain over 90 days. 1% for the plain pack (603 orders). 4 points. The direction is credible; the exact size is not.

Write the assumptions beside the number: the window, the sample, the exclusion rules for returned orders. A model nobody can reconstruct six months later is not evidence, it is a memory.

Three failures that waste a test

Three failures recur: the first is measuring everything and reporting nothing, which yields a forty-tile dashboard with no decision.

The second is comparing this quarter's packaging test with last quarter's baseline despite a pricing change in between. The third is treating an unboxing post spike as proof of sales impact without checking if sales moved.

A short audit of packaging marketing measurement mistakes covers the traps in more detail, including seasonal distortion around gifting periods. Fixing attribution errors after a budget review is expensive. Fixing them before is a spreadsheet exercise.

One more discipline: record the cost of measurement itself. Surveys, holdout logistics and analyst time are real costs, and a test that costs more than the format change it justifies is not worth running. Put that figure in the same table as the result.

Report so the numbers get used

A report exists to trigger a decision, not to archive activity. If a reader cannot say what changes as a result of reading it, the report is too long or too vague.

Keep the front page to six measures with a comparison against the previous period and the same period last year. Put method notes on page two. Put raw data in an appendix that only analysts open.

One monthly dashboard, named owners

Most teams already have order data, review data and post metrics in separate tools. The reporting job is to bring the packaging-specific measures alongside them without creating a parallel universe of numbers.

A single recurring dashboard, refreshed monthly, with named owners for each metric, removes the argument about whose figure is correct. It also makes the quarterly review fast, because the audience has seen the trend already.

A metric without a written definition will be calculated two ways within a quarter. Define repeat purchase as orders within 90 days of the first order, state whether returns are excluded, and put that definition next to the number. The habit prevents most internal disputes before they start.

Our walkthrough of how to structure a packaging marketing reporting dashboard sets out a layout that works for both marketing and operations readers. Include a short written commentary each month. Numbers without a sentence of interpretation get misread, especially by people outside the team.

Reporting cadence matters as much as layout. Monthly is right for operational measures such as damage rates. Quarterly suits brand measures such as repeat purchase, which move slowly and need a longer window to mean anything.

Tie the report to planning and risk

Packaging budgets are approved alongside wider commercial plans, so the report should speak the language of the planning cycle. That means connecting packaging measures to demand context. The Office for National Statistics publishes output and productivity data that helps frame whether a soft quarter is your packaging or the wider market, at ONS economic output and productivity statistics.

Risk belongs in the same conversation. Customer data captured through QR codes, inserts and apps sits under data protection rules, and the security expectations around those systems are set out by the BSI at cybersecurity standards guidance. If your measurement plan involves tracking scans back to individuals, that is a compliance question before it is an analytics one.

Planning also needs a forward view. Recycled-content requirements, tax thresholds and consumer expectations all move, and our packaging marketing trends and outlook for England sets out what to watch in the 2027 planning round.

Benchmark your engagement assumptions against wider marketing research rather than internal folklore. The Data & Marketing Association publishes industry research on customer engagement that is useful for sanity-checking response rates at DMA research. If your insert redemption rate is ten times the industry norm, check the measurement before you celebrate.

Glossary

Common questions

How long should a packaging test run?

Long enough for repeat behaviour to show, which usually means 90 days for consumables and longer for gifting. If your order volume is low, extend the window rather than widening the claim.

Can we measure packaging without a holdout?

Yes, but be honest about the limits. Cohort comparison and self-reported attribution are workable substitutes, provided you label the result as indicative rather than causal.

How often should the report change?

Layout should stay stable for at least a year so trends remain readable. Metrics can change annually at planning, with the reason recorded in the commentary.