Content marketing: a data-led guide for 2026

Content marketing is in an awkward moment. The tactics that worked from 2015 to 2022 — publish frequently, cover every question, wait for compounding traffic — have been undercut by AI answers absorbing exactly the kind of query those tactics targeted.

That does not make content marketing obsolete. It makes the undifferentiated version obsolete, which is most of it.

What content marketing is for

It earns attention by being useful before you ask for anything. In practice it does three jobs:

• Discovery. Being found by people researching a problem you solve.
• Credibility. Demonstrating competence rather than claiming it.
• Sales support. Answering objections in a form that can be sent, read and referred back to.

Most programmes optimise only for the first, which is also the job AI answers have made hardest. The second and third have become relatively more valuable.

What has actually changed

Informational queries return less. Questions with a short factual answer increasingly resolve in a summary. Building a strategy on them buys less than it used to.

Citation has become a channel. When an assistant answers a question by quoting three sources, being one of them is a form of visibility that did not exist five years ago and is not measured by traffic. See generative engine optimisation.

The floor has risen. Producing competent-sounding text costs almost nothing now, so competent-sounding text is worth almost nothing. What is scarce is specificity: real numbers, real cases, real method.

The formats that still earn their cost

Measuring return

Work backwards from revenue:

1. Organic sessions to the page
2. Conversions attributable to it, including assisted
3. Lead-to-customer rate for that source
4. Average deal value
5. Cost to produce and maintain

The two commonly missed items are assisted conversions — guides often contribute to deals they do not close — and maintenance, since content decays and the refresh is part of the cost.

If you cannot attribute organic conversions at all, fix that first; see our Google Analytics 4 guide.

Why content marketing fails

In roughly descending order of frequency:

• No commercial anchor. Traffic arrives with nowhere to convert.
• Publishing to a calendar. Topics chosen by deadline rather than by value.
• Insufficient depth. Enough to publish, not enough to rank.
• No distribution. Publishing is not promotion; a good piece nobody sees has failed.
• Abandonment. Stopping at month four, one month before the compounding would have started.
• No refresh. Treating published as finished.

What it costs in the UK

The main cost is time from someone who knows the subject. A substantial, well-researched guide takes one to three days of expert-informed work including research, drafting, review and internal linking. Price that at whatever senior time costs in your market and the arithmetic is straightforward.

Outsourced, the range is wide: a few hundred pounds per piece for competent generalist writing, considerably more where subject expertise, original data or regulatory review is involved. The cheap end is rarely worth buying now, because it competes directly with free AI output and loses.

Scaling internationally

Translation is the easy part and the least valuable. What matters is that search behaviour differs by market: different phrasings, different competitors, different intent distribution for what looks like the same term.

A translated page targeting the wrong local query ranks for nothing. Do the keyword research per market, then translate against those findings — which is also why this English site exists separately rather than as a machine translation of the Finnish one.