Digital Marketing Is Two Thirds of All Ad Spend. Here Is How to Measure It

The four numbers that decide a budget, and the blind spot that now hides most of the buyer journey

Digital marketing is every paid, earned and owned activity that reaches someone through a connected screen, and in 2026 it carries about 69% of all advertising worldwide. You measure it with four numbers: acquisition cost, return on ad spend, conversion rate and lifetime value. The difficulty is that much of the influence now happens where no analytics tag can follow.

Table of Contents

1. The definition, and the half that gets skipped
2. Where the money actually goes
3. The four numbers that run a digital marketing budget
4. Why your dashboard says Direct
5. How to measure what the tags cannot see
6. A thirty day setup that ends in a holdout

The definition, and the half that gets skipped

Most explainers define the category and stop at a list of channels: search, social, email, display, streaming video, retail sites. That list is accurate and almost useless, because it tells a reader what exists without saying what any of it is for or whether it paid back. The spend says the stakes are no longer small: of roughly $1.30 trillion spent on advertising worldwide this year, about $836 billion is digital.

That share of global ad investment has climbed from about 64% in 2021 to roughly 69% today, which changes the beginner question. It is no longer whether to advertise on a screen; it is how to read the result when almost every channel is a screen. Everything below is the second half of the answer, the measurement half, written for someone who has opened an analytics dashboard and not quite trusted it.

Where the money actually goes

Inside digital ad spend the split across channels is not what the advice implies. Search takes about 40% and social about 32%, display roughly 18% and video near 10%. Search leads for a plain reason: somebody typing a query has already decided they want something, so the advertiser is buying a place in front of an intention that already exists.

Social sits second because it buys attention before any intention has formed. The fastest growing line items are newer than either: retail media, the sponsored placements inside retailer apps, and connected TV, which is advertising inside streaming apps, each growing around 14% a year. Both sit close to a purchase, and both are the hardest to trace back to one, which becomes the problem further down this page.

The four numbers that run a digital marketing budget

Twenty metrics on a dashboard is not measurement. Four numbers carry the decision, and each answers a different question: what a customer costs to acquire, what each advertising dollar returns, what share of visitors convert, and what a customer is worth across their whole life with you. Everything else, impressions and reach and engagement rate, is a diagnostic that explains those four rather than a result.

• Acquisition cost is total spend divided by new customers, not leads. It is the price of growth.
• Return on ad spend is the revenue credited to a campaign divided by what the campaign cost, read per campaign and per channel. One company wide figure hides every decision worth making.
• Conversion rate is the share of visitors who do the thing you want, and it is the cheapest lever on the list because it multiplies every channel at once.
• Lifetime value is the gross profit a customer produces before they leave you.

The decision lives in the ratio between the first number and the last. Acquisition cost below a third of lifetime value means there is room to spend harder; acquisition cost above lifetime value means every new customer makes the company poorer, however good the click through rate looks on the slide. There are longer lists of metrics worth tracking, but these four are what a budget review actually turns on.

Why your dashboard says Direct

All of that assumes you can see where a visit came from. When a browser follows a link it tells the destination which page sent the visitor, a value called the referrer, and AI assistants and private messaging apps send nothing. So a buyer who asks an assistant to compare three products, reads the answer, then types your brand name into the address bar arrives labelled Direct, which is the label analytics uses when it has no idea.

The scale is not marginal. In one measured sample about 71% of confirmed assistant referred sessions arrived with no referrer, and Similarweb research found assistant referred traffic growing 393% year over year while converting about 42% better than everything else. Links passed around in Slack, Discord and WhatsApp behave the same way, the older and quieter version of the same dark funnel.

How to measure what the tags cannot see

You can still measure this work. You stop expecting one dashboard to be the whole answer and start bounding the truth from three sides. The first is to treat Direct as a channel under investigation rather than as loyalty: if Direct rises while branded search rises and paid clicks stay flat, something off platform is creating demand.

Add one question to the signup form asking where people heard about you, then compare those answers against what your attribution model claims; the gap is the size of the blind spot. Benchmark against the category rather than against last month, since a flat quarter may belong to the market. Then ask what your brand looks like inside AI answers, because the assistant is now a channel.

A thirty day setup that ends in a holdout

The strongest measurement available to a small team costs nothing except nerve: switch a channel off in one region, keep it running in a comparable one, then compare total revenue rather than channel attributed revenue. That difference is incrementality, the sales a channel caused rather than the sales it merely stood beside. Teams who run it on branded search are regularly surprised by how much of the credited revenue arrives anyway.

1. Week one: define one conversion event and one revenue figure, and get finance to agree to both.
2. Week two: tag campaigns consistently and add the source question to every form.
3. Week three: build one view holding the four numbers, split by the job each channel does rather than by platform.
4. Week four: start a single holdout in whichever channel takes the largest share of spend.

At the end of it you will not have perfect attribution, and nobody has that. You will have a defensible answer about what to fund next quarter, which is the thing the perfect dashboard was only ever a proxy for.

Questions worth answering first

What is the difference between digital marketing and traditional advertising?

Delivery and measurement. A connected screen records most of what it delivers, while print, outdoor and broadcast reach people without a per person log. The gap narrows from both ends, since connected TV behaves like broadcast with an ad server attached.

Which channel should a small team start with?

Start with capture, because it is cheaper per sale and the payback is quick: search advertising, the organic results above it, and retail media at the digital shelf. Add creation work once capture plateaus, which it will, since you cannot capture more demand than exists.

What counts as a good return on ad spend?

There is no universal figure, because it depends on gross margin and on how much of that revenue would have arrived anyway. The usable version is the ratio of acquisition cost to lifetime value: under a third is room to spend, above one is a leak.

Why is Direct traffic growing in my analytics?

Usually because the referrer is missing, not because loyalty is rising. Assistants, private messages and community apps send no referrer, so those sessions land in the Direct bucket by default. Check whether branded search is climbing at the same time, which is the tell.

This article expands on our video “What is digital marketing and how do you measure it?”. Watch it here.