Analytics and Tracking

Calculated Field

Also called Custom field, derived metric

A metric or dimension defined by formula inside a report, worked out from existing fields each time it is viewed.

Quick facts: Calculated Field

Category
Analytics and Tracking
Also called
Custom field, derived metric
Level
Intermediate
Affects
Report clarity, metric definitions, decision quality
Where to see it
Looker Studio, GA4 calculated metrics, spreadsheet formulas
In this article4
  1. How a calculated field works
  2. Why calculated fields matter
  3. Common mistakes with calculated fields
  4. How to act on it

How a calculated field works

You write a formula using fields the source already provides, and the reporting tool works it out whenever someone looks at the chart. Nothing is written back to the platform: the field exists only inside the report, and it recalculates against whatever date range and filters the reader has selected.

There are two useful kinds. A derived metric combines numbers — spend divided by enquiries to get a cost per enquiry, revenue divided by spend, a total built by adding several event counts together. A derived dimension reshapes labels instead — grouping a long list of campaign names into channels, tidying inconsistent country spellings, or extracting the section of a website from its address so pages can be grouped by folder.

The subtlety that catches people is when the calculation happens. The tool aggregates the underlying fields for the rows on screen, then applies your formula to the aggregate. Change the date range or remove a filter and the answer legitimately changes, because the inputs did.

Why calculated fields matter

They let a report speak the business’s language rather than the platform’s. A shop does not think in link clicks; it thinks in cost per order. A consultancy does not think in form submissions; it thinks in qualified enquiries, which may be a subset of them. Without calculated fields the reader has to do arithmetic in their head while looking at the chart, and they will get it wrong.

They also rescue messy data. Years of inconsistent campaign naming can be grouped into something readable by one derived dimension, without touching the accounts themselves.

Common mistakes with calculated fields

Averaging a ratio is the one that produces genuinely false numbers. Working out a cost per enquiry for each day and then averaging those daily figures gives a different answer from dividing total spend by total enquiries, and only the second is correct. Always build ratios from summed totals.

Dividing by something that can be zero is the next, which turns a chart into a row of errors on any day with no conversions. Rebuilding the same formula on every chart is the quiet one: the copies drift as people edit them, and eventually two charts on the same page disagree. Building a calculation on top of a blended data source compounds whatever the blend got wrong.

How to act on it

Define each field once at source level in the report and reuse it, so one edit updates every chart. Name it so the definition is obvious — writing which conversions are included is more useful than a clever short name — and put the wording somewhere the reader can see, because a metric nobody can define is a metric nobody trusts.

Build ratios from totals, guard against empty denominators, and check each new field against a manual calculation before it goes anywhere near a client report. If a definition matters to how the business is run, agree it with whoever makes the decisions rather than inventing it inside a reporting dashboard and hoping it matches their spreadsheet.

Do and do not

Do

  • Define each field once and reuse it everywhere
  • Build ratios from summed totals, never from row averages
  • Write the definition where the reader can see it

Do not

  • Average a ratio across days
  • Divide by a field that can be zero
  • Hide a tracking fault behind a formula

Questions people ask about this

Why does my cost per conversion change when I change the date range?

Because a calculated field is worked out from whatever rows are currently in view. Spend and conversions are both totalled for the selected period, then divided. A different period has different totals, so the ratio moves. That is correct behaviour. What would be wrong is a figure that stayed the same regardless of the period selected.

Should I calculate metrics in the report or in the platform?

Calculate in the report when the definition is yours and may change, such as grouping campaigns into channels or excluding certain conversion actions. Fix it at the platform when the underlying data is genuinely wrong, for example when conversions are being counted twice. Patching a tracking fault with a formula hides the fault and eventually somebody trusts the raw number.

Can a calculated field change the data in Google Ads or GA4?

No. It exists only inside the report and is worked out each time the page loads. The platform never sees it, and nothing you write in a formula can alter the underlying account. That makes calculated fields safe to experiment with, but it also means the definition lives in the report and must be documented there.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.