office, business, notebook, statistics, technology, digital, computer, internet, marketing, accounting, corporate, job, communication, creative, table, meeting, businessman, ideas, woman, project, coworking, brainstorming, loft, plan, documents, discussion, group, man, team, mobile, mock, work, laptop, professional, startup, management, teamwork, network, account, paperwork, hipster, young, cooperation, object, papers, global, millennials, living room, sofa, blue business, blue computer, blue office, blue technology, blue laptop, blue work, blue meeting, blue paper, blue marketing, blue team, blue mobile, blue network, blue community, blue room, blue internet, blue digital, blue table, blue group, blue job, blue company, blue conference, blue global, blue businessman, blue teamwork, blue plan, blue planning, blue creative, blue document, blue accounting, blue idea, blue management, statistics, statistics, marketing, marketing, marketing, accounting, accounting, accounting, accounting, accounting, project, management
Photo by Goumbik on Pixabay

Measurement

Part of What shopping feeds measurement means for multichannel reporting

Shopping feeds benchmark research or a gut-feel budget call

A practical route through shopping feeds benchmark research: named UK sources, source quality checks, shipping data, the calls where judgement beats research.

Editorial Desk · · 5 min read

What to take away

  • A Manchester retailer comparing its return rate with a trade body average is doing benchmark research, even without the label.
  • Benchmarks need a named source, a date and a defined population before they mean anything for a budget line.
  • Shipping settings sit inside the benchmark, because Google Merchant Centre treats delivery data as part of the listing itself.
  • Shopping feeds budgets are usually small enough to test and change back, which is exactly when judgement is allowed to stand in for research.
  • Two or three external anchors beat a long list of numbers nobody can trace.

What benchmark research actually is

A benchmark states who was measured, over what period and by which method. Without those three details, it is an anecdote wearing a suit.

Benchmark essentials

  • Who was measured
  • Over what period
  • By which method
  • Metriccost per click
  • Metricreturn rate
  • Metricfeed error counts
  • Metricorder value

For feed work, benchmarks usually cover cost per click, return rate, feed error counts or order value. Each one needs a population.

Start by writing the claim you want to test. Then find the smallest credible source that answers it, and tie that claim to a reporting line before you add another column to a spreadsheet.

Source quality checks

Not every published number is a benchmark. Some are marketing. Check the publisher first, then the method note.

Starting points that carry a publisher: the ONS retail sales figures for national context, IMRG for UK online retail trade benchmarks, and the reporting already sitting in your own Google Merchant Centre and Google Ads accounts for account-level cost per click, return rate and feed error counts. A trade body figure, a government statistical release and your own account history are the three sources most budget cases need, and each one should be recorded with its date.

Government and regulator pages are strong for legal context, not for conversion rates. The ICO explains the Data (Use and Access) Act 2025 and what it changes for organisations handling personal data. That matters when your benchmark uses customer-level order records rather than aggregate totals.

For claims about prices, delivery promises or savings, the Consumer Protection from Unfair Trading Regulations 2008 set the standard for what a retailer can state. A benchmark that contradicts your own published claim is a compliance risk, not a stretch target.

Shipping data in the baseline

Delivery cost and speed change the comparison more than most teams expect. Google Merchant Centre asks sellers to add shipping information because the listing is judged on the total offer, not the product alone. The Merchant Centre shipping information guidance explains what to supply.

If your benchmark uses a competitor's headline price, check whether delivery is included. A free-delivery threshold can move the effective price by several pounds. For example, a shop charging £4.95 on a £30 basket is quoting a 16.5 per cent uplift that a like-for-like comparison must recognise.

A numbered route to a defensible benchmark

  1. Write the decision the benchmark will inform, such as whether to raise bids on one category.
  2. Define the populationyour own account, your category, or a published national figure.
  3. Find the source and record its publication date.
  4. Normalise the units, so you compare cost per click with cost per click.
  5. Adjust for shipping, tax and returns before comparing.
  6. Store the figure with its source in a reporting table.
  7. Review the benchmark at a fixed interval and retire stale numbers.

A number without its source cannot be defended in a budget meeting, so keep the publisher, the date and the population in the same row as the metric it explains.

When gut feel is the better call

Benchmark research costs time, and that cost has to be weighed against the size of the decision and how easily it can be undone. A small, reversible tweak, such as testing a lower bid on one category for a week, can be decided on judgement and corrected if it goes wrong.

A benchmark earns its place when the decision is expensive, hard to reverse, or likely to be challenged later — changing the delivery promise across a whole catalogue, or committing to a new feed platform.

Work through these tests before you decide:

  • Size of the commitmentthe more budget the decision ties up, the more a benchmark is worth the effort.
  • Reversibilitya setting you can change back next week needs less evidence than an annual contract.
  • Cost of the researchif finding and normalising a source takes longer than the decision is worth, decide on judgement.
  • Challengeif a finance or compliance review will ask where the number came from, bring the named source.
  • Fit of the populationif nothing measures a category close to yours, use your own account history or decide on judgement.

Record the reasoning either way. A judgement call written down with its context is defensible; an unrecorded one is not.

Reporting the result

Report the range, not the point estimate. If three sources give return rates between four and nine per cent, say so. A single figure invites false precision.

Label every illustrative number as an example. A team paying £400 a month for feed software is an example, not a market rate. That line stands for a subscription of the kind sold by feed management platforms such as Channable, DataFeedWatch or Feedonomics, so check current pricing with the supplier rather than treating the figure as a rate.

Common questions

How many sources does a benchmark need?

Two or three credible sources are usually enough for a working comparison. More sources add noise unless they measure the same population.

Can I use my own account history as a benchmark?

Yes, provided you note the period and any changes to feed rules or pricing. Your own history is often the most relevant comparator available.

Where should benchmarks live?

In the same reporting table as the metric they support, with a source column. Separate documents get lost during peak trading.

Do benchmarks differ across the UK?

Published national figures usually cover the whole UK, while delivery costs and competition vary by region. State which population you are using rather than assuming one figure fits every market.

Filed undershopping feedsbenchmark researchmultichannel ecommercemeasurementproduct feed software

More in Measurement