The Ecommerce KPI Scorecard: Which Numbers, Which Definitions, Which Targets

A KPI scorecard is short on purpose: around ten numbers, each with one written definition, a target for the period, and a person who answers for it. Picking the numbers takes an hour. Agreeing on what each one means takes the rest of the month, and that agreement is the part that makes the scorecard worth opening. A scorecard where revenue means one thing to marketing and another to finance produces an argument every Monday and no decisions.

This guide covers which numbers belong on a webshop scorecard, how to define each so that everyone means the same thing, how to set them against a target or a budget, and how a scorecard differs from a dashboard. It works in any tool, from a spreadsheet to a BI suite.

What a KPI scorecard is

A KPI scorecard is a single view that shows, for a fixed set of key performance indicators, the actual value for a period next to the target for that period, with a status that tells you whether you are on track. It answers one question: are we where we said we would be?

It is the smallest working form of business performance management. Business performance management is the practice of setting goals for a company, measuring actual results against those goals at a fixed rhythm, and adjusting plans or actions based on the difference. The scorecard is where that comparison becomes visible. Without targets it is a list of numbers. Without a fixed rhythm it is a report that people stop opening.

How a scorecard differs from a dashboard

The two get confused because they often sit in the same tool and share the same data. They do different jobs.

A red number on the scorecard is the start of a conversation. The dashboard, or a question asked in plain language, is where you find out why. Both should read from the same definitions. If net revenue on the scorecard differs from net revenue on the channel dashboard, people stop trusting both within a week.

The numbers that belong on a webshop scorecard

There is no universal list, but most webshops end up close to this one. Start here and remove what you would not act on.

Treat ten as a ceiling. If a number has been green for a year and nobody would react if it turned red, it does not belong on the scorecard. Move it to a dashboard.

One opinion from experience: put margin after marketing on the scorecard even if the first version is rough. A shop can hit its revenue target and its return on ad spend target in the same month and still earn less than the month before, because the growth came from discounted orders with a high return rate. Revenue and spend alone will not show that.

Define every number once

Here is one month for a hypothetical webshop, written the way the money actually moved.

Four people can each report "revenue" from this table and all be right: €261,360, €216,000 with shipping, €210,000 or €189,000. The highest is 38 percent above the lowest. With €30,000 of ad spend, the marketing cost ratio is 11.5 percent on one and 15.9 percent on the other. A target of 14 percent is met or missed depending on which number someone pulled.

So each number on the scorecard gets a short definition card. Six lines are enough:

1. Name. One name, used everywhere. If the scorecard says net revenue, no report says turnover.
2. One sentence in plain language. "Net revenue is what customers paid for products, after discounts and returns, without VAT and without shipping."
3. What is in and what is out. Gift cards, wholesale orders, test orders, marketplace sales, cancelled orders.
4. Time basis. Order date or payment date, and which time zone closes the day.
5. Source of truth. The system that wins when two systems disagree.
6. Owner. The person who decides when the definition has to change.

Three choices cause most of the trouble.

VAT. Leave it out. VAT is money you collect for the tax authority, rates differ per country and per product group, and a number that includes it moves when your country mix moves. The store's own reports often show totals including tax, which is why this needs to be written down.

Discounts and returns. Take both off. A scorecard that shows revenue before discounts rewards the team for discounting. Returns need a second decision: do they count against the month the order was placed or the month the product came back? Counting them in the order month gives the truest picture of that month but means recent months keep changing for a few weeks. Counting them in the return month gives stable numbers that tie to the bookkeeping. Finance usually prefers the second. Choose one and say so on the card.

New versus returning customers. The usual definition is a first order ever, matched on email address. Decide what happens with guest checkouts, with a customer whose only earlier order was fully refunded, and with someone who comes back after three years. Ad platforms call many of these people new. Your scorecard should use your own order history, because that is the only source that knows who bought before. Why the platforms and the store disagree in the first place is covered in Why Shopify, GA4 and Meta All Report Different Revenue.

Set every number against a target

A number without a target cannot be red or green. Three kinds of reference value are useful, and they are not the same thing.

• Target. What you aim to reach: €2.6 million net revenue this year.
• Budget. What you plan to spend: €30,000 of ad spend in October.
• Guardrail. A limit you do not want to cross: a marketing cost ratio above 16 percent, a return rate above 12 percent.

Each number also needs a direction. Higher is better for revenue and worse for return rate. Tools that color a number green because it went up, whatever the number is, cause more confusion than they prevent.

Then spread the yearly figure over the months the way your business really sells. This is where most scorecards go wrong. Say November has a net revenue target of €260,000. On the thirteenth, a straight line says you should be at €112,667. But a shop that makes a third of its November revenue in the last week should be at roughly €98,000 by then. Measure against the straight line and the scorecard shows red for three weeks, then green on the last day. People learn to ignore it. Use last year's daily pattern to spread the month, and the status means something from day one.

Finally, agree on what the colors mean in numbers. For example: green within 3 percent of plan to date, amber within 10 percent, red beyond that. Write the thresholds down next to the definitions.

Build it in the tool you already have

A scorecard needs four things from a tool: the actual value from an agreed definition, a target stored per metric and per period, the difference between the two, and a status. A spreadsheet can do it, and so can Data Studio (the former Looker Studio), Metabase, Power BI or any other BI tool. The tool matters less than two habits.

Keep targets as data. Store them in one table with a row per metric and per month, next to the actuals. Do not type a target into the settings of a single chart. A typed target exists in one place, cannot be reused, and is still showing last year's figure in February.

Compute the number once. If the scorecard calculates net revenue in one place and the channel dashboard calculates it again in its own formula, they will drift. Define it once, upstream of both, and let every view read the result. The same goes for an AI assistant: it should read that definition and not work out its own, which is the subject of MCP for Your BI Data.

Where the data for the actuals comes from, and how it reaches your reporting tool, is covered in the guide to getting marketing and webshop data into your BI tool.

Review it at a fixed rhythm

Weekly works for most webshops, in a meeting of thirty minutes or less. Go through the numbers that are off plan and skip the rest. For each one, name who looks into it and by when. The investigation itself happens outside the meeting, on a dashboard or in a conversation with the data.

Change the scorecard itself once or twice a year. If the list changes every month, nobody builds a feel for the numbers, and comparisons with earlier periods stop working.

Where Verity fits

Verity keeps this in one place for marketing and ecommerce data. Each metric has one approved definition, explained in plain language in a business glossary, and targets, budgets and guardrails are stored per metric and per period. A dashboard, a scheduled report and an answer from an AI assistant all read that same definition, so the number on the scorecard is the number everywhere else. Monitors can watch a number against its target and tell you when it moves. The overview is on the AI analytics and dashboards page.

Frequently asked questions

What is the difference between a KPI and a metric?

Every KPI is a metric, but few metrics are KPIs. A metric is anything you can count. A KPI is a metric you have set a target for and would act on if it went off plan. Click through rate is a metric. Margin after marketing against this month's plan is a KPI.

How many KPIs should a scorecard have?

Around ten for the company, and fewer is fine. Every extra number makes it less likely that the important ones get discussed. Teams can keep their own, more detailed dashboards underneath.

Should revenue on the scorecard include VAT?

No. Report revenue without VAT, after discounts and returns. That is the number that can be compared across countries and months and that ties to the bookkeeping. If the store's own reports show totals including tax, note the difference on the definition card.

What is the definition of business performance management?

Business performance management is the practice of setting goals, measuring actual results against them at a fixed rhythm, and adjusting plans or actions based on the difference. A KPI scorecard is the view where that comparison is made.

How often should targets change?

Set yearly targets once and spread them over the months. Revise during the year only when the plan itself changes, for example after a new market launch, and keep the old target visible so the history stays honest.