The digital marketing KPIs worth reporting, and the ones to stop

The digital marketing KPIs worth a monthly report come down to six things, and a firm can hold all six in its head.

• The one outcome you agreed at the start, and where it stands.
• Qualified enquiries in the period, against the definition you wrote down.
• What happened to them: contacted, met, proposed, won, lost, and why.
• What one qualified enquiry cost.
• Search and site visibility on the terms with commercial intent, not all terms.
• What changed this month, what you learned, and what you are doing differently next month.

Everything else is a diagnostic. Diagnostics matter when something breaks: click-through rate tells you the ad is wrong, bounce rate tells you the page is wrong, frequency tells you the audience is too small. They are tools for the people doing the work, not headlines for the people paying for it.

Most marketing reports invert that. Page one carries reach, impressions, followers and engagement, and the commercial question appears on page nine if it appears at all. That is what happens when a supplier is accountable for producing outputs: the report becomes evidence that outputs were produced. The reporting chapter sets out what we send and when.

A KPI is a key performance indicator, which in practice means a number tied to a decision. There is a simple test for whether something qualifies. If this number halved next month, what would we do differently? If the honest answer is nothing, it is not a KPI. It is a fact about the month.

Apply that test to a typical report and most of it falls away. Followers halve: you would do nothing, because followers were never the route to a client. Impressions halve: you would check whether spend dropped, then do nothing. Qualified enquiries halve: you would stop the meeting you were in and find out why.

Three or four KPIs is the right number for a board or an owner. Below that you are flying blind. Above that nobody remembers any of them, and the report becomes a document to be received rather than read.

If this number halved next month and you would do nothing, it is not a KPI. It is a fact about the month.

CVR in digital marketing means conversion rate: the share of a defined group of people who did the thing you asked, inside a defined window. Visitors who enquired. Enquiries that became meetings. Meetings that became clients. It is the most useful ratio in marketing and the easiest to misread, because a ratio hides both numbers that made it.

Two traps. The first is the denominator. Buy cheaper traffic and conversion rate falls while enquiries rise. Narrow the audience to people already looking for you and conversion rate climbs while the business gets smaller. The rate alone will not tell you which happened.

The second is that conversion rate improves when you ask for less. Strip the fields from a form and more people complete it, fewer of them are worth calling, and the report looks better until sales sees the list. That trade sits behind a lot of campaigns that hit target and delivered nothing, which is the argument in lead generation is not the job.

So report the two raw counts beside the rate, always, and name the window and the population. And ignore any industry benchmark for a good conversion rate. Your own previous quarter, measured the same way, is the only comparison with meaning in it.

None of these are lies. They are inputs, presented as results, and the tell is that every one of them can rise in a month when nothing commercial happened.

• Impressions and reach. A measure of how much media was bought, not of whether anyone cared.
• Followers. A number that grows with time and posting, and that a competitor can buy in an afternoon.
• Engagement rate. Useful to the person choosing what to post next. Meaningless as a business indicator.
• Posts published, articles written, hours worked. A record of activity, which is what you were trying to avoid paying for.
• Keyword positions in bulk. Two hundred terms improved matters far less than the six with commercial intent.
• Platform-reported conversions, quoted without saying they are platform-reported.

We have our own evidence for how little output counts prove, and it sits in how to choose a marketing agency. The point for a report is narrower. Every monthly report through that period would have shown articles published, on schedule, as promised. Not one of those reports would have been inaccurate, and none of them would have told you anything.

You usually cannot measure the return on social media marketing the way the question implies, and a supplier who says otherwise is quoting a platform back to you.

Buyers see a post, say nothing, and forward it to a colleague on WhatsApp. B2B decisions take months and several people, and the last click gets the credit for all of it. Platforms count view-through and modelled conversions inside their own walls, so three of them will each claim the same client. Privacy changes have broken the chain between a view and a purchase.

What you can do instead is more useful than a false number.

• Ask every enquirer how they heard about you and record it in the CRM. It is imperfect, and it beats everything else.
• Test by switching off. Hold a campaign or a market dark for a defined period and watch what happens to enquiries. Crude, and the closest thing to proof available.
• Judge organic social on the job it can do: keeping you present with people who already know you, so you are the name they think of when the need appears.

The wider argument, that social is one tool inside marketing rather than the whole of it, is in social media is not marketing.

Email is the one channel where you own the audience, so the metrics deserve better treatment than they usually get.

Open rate is no longer a measurement. Privacy features that pre-fetch images register opens that never happened, and the figure moves with the mail clients your list happens to use. Watch it as a rough trend across your own sends and never report it as a result.

What to track instead

• Clicks to the page that matters, not clicks to anything.
• Replies. In B2B this is the strongest signal an email list produces.
• Unsubscribes and spam complaints, read together. A rising complaint rate is the early warning before deliverability goes.
• Deliverability itself: are you reaching the inbox, and for which domains.
• List growth by source, kept separate. A list bought or scraped is a liability that will eventually damage the sending reputation of the list you built.

Then segment everything by how the address arrived. An enquiry form, a white paper download and a conference badge scan behave nothing alike, and averaging them produces a number that describes nobody.

Analytics tells you what happened on property you own, in aggregate, with gaps. It is very good at that. It cannot tell you why anyone did it, it cannot see the message forwarded on WhatsApp, and it cannot connect a signature this quarter to the webinar someone attended last year.

Which means the work is narrower than most setups assume. Get four things right and stop.

1. Every enquiry route is recorded as an event, including the WhatsApp click and the phone number tap.
2. The source of each enquiry lands on the CRM record and survives to the point where the deal is won or lost.
3. One source of truth for the numbers that leave the building, agreed in advance, so nobody arrives at the meeting with a different figure.
4. Self-reported attribution, that single question on the form, captured every time.

After that, more instrumentation mostly produces more dashboards. A dashboard nobody has ever changed a decision because of is a cost, not an asset.

One page. The outcome you agreed and where it stands. Qualified enquiries and what happened to each of them. What one cost. The pipeline created and its state. What you changed this month and what you learned. What changes next month, and what you have stopped doing.

That is the whole report. If it runs to forty slides, either it is written for a different audience or it is hiding something, and the most common thing it is hiding is that nobody agreed what success was before the work began.

The test comes at the end of the year. When your board sets next year's budget, marketing appears as a line like every other cost, and someone asks what it did. For most firms the honest answer is a set of numbers about attention, and the line survives because cutting it feels like giving up.

The alternative is a line nobody argues about, because everyone in the room can name the problem it solved. That is the only standard worth holding a marketing department to, ours included.