PlaybookRU

Playbook8 min read

How to measure content marketing ROI when attribution is broken

Your analytics will never show every visit that led to a sale, so the honest answer to content ROI is a range built from three numbers you can check. This article shows how to build that range and when to distrust it.

A hand points to colorful charts on a sheet of paper
A chart is only as useful as the question you bring to it

Can you prove what a blog earned if the reports are incomplete?

Not to the dollar, and anyone who claims it is rounding. You can prove a range: a floor made of deals your tracking can follow, and a ceiling made of deals where the buyer demonstrably read your content at some point. The gap between the two numbers is your uncertainty, and stating it openly is what makes the rest of the report believable.

This matters because the standard report was built for paid ads. A click on an ad, a form, a sale. Content works on a different clock. A buyer reads a comparison page in March, forwards it to a partner, returns by typing your name in September and fills in a form. Last-click credits September's visit and nothing else. One practitioner write-up found single-touch models crediting content with about 8% of pipeline, against about 27% when the same deals were re-cut with multi-touch market. Reviewing those reports also takes your time, so check how many hours a month your agency needs.

That is one team's audit. It shows the direction of the error, undercounting, and says nothing about the size for your site. Set the result against the invoice with the SEO retainer cost lines in hand.

Where does the credit go missing?

Three places account for most of the loss, and each one has a different fix.

Private sharing. Links pasted into messengers, team chats and email carry no referrer, so the visit lands in "direct". The same buyer-committee dynamic is described in B2B attribution write-ups on dark social market. You cannot recover the lost touch from analytics, so the fix sits in the form: ask the buyer.

AI answers. A reader who first met your brand inside an assistant's answer often arrives later by typing the domain. The ranking report shows no click at all, and the sale gets filed under direct traffic. Our look at what changed now that AI answers first covers how to read a blog in that situation.

Long gaps. If your average sale takes three months from first visit to signature, a 30-day tracking window drops the first touch from most deals by construction estimate.

Colorful banknotes lying on top of printed financial charts
Cash on a chart is the number everyone wants and the one analytics cannot hand over

What is the cheapest fix for missing attribution?

A single mandatory question on every contact form and every call intake: "How did you hear about us?" Make it free text, since people write things a dropdown never lists, such as "ChatGPT suggested you", "a colleague sent your pricing article" or "saw your post and came back later".

Read the answers monthly and sort them into plain buckets: search, AI assistant, referral, social, "your blog". Compare the count with the analytics report for the same month. The disagreement is the finding, because it sizes the hidden share for your own site.

Self-reported answers lean toward the last thing a buyer remembers, so treat them as a second instrument next to analytics. When both point the same way, the conclusion is safer than either alone.

A three-line ledger you can fill in today

Keep one table per quarter and fill it from your CRM. Every row below is a count you can check by hand.

LineWhat to countExample quarter
CostWriters, editing, tools, your own time at a fair rate$9,000 estimate
FloorRevenue from deals whose tracked path includes a blog visit$14,000 estimate
CeilingRevenue from deals where the buyer read the blog at any point, per form answers or CRM notes$38,000 estimate
Итого rangeFloor to ceiling, divided by cost1.6x to 4.2x estimate

In this example the blog paid back between 1.6 and 4.2 times its cost, and both ends clear break-even estimate. A floor of 0.5 and a ceiling of 1.1 would say the blog is not clearly paying and needs a harder look at which topics bring buyers estimate.

The ceiling overstates the effect, since a buyer who reads a post and then signs with you may have signed anyway. The floor understates it for the reasons in the previous sections. Report both, and say which one you used for a given decision.

The same ledger works when an agency produces the content: ask for the cost line itemized and keep the revenue lines on your side. If you are still weighing who should write, how to choose an SEO agency lists the questions that surface reporting habits early.

Why a longer window changes the verdict

Content costs money at publication and pays later, so a short window compares full cost against partial return. Score each article against the quarter it was published in and almost every post looks like a loss.

Use two windows instead. The first is a rolling year for the whole blog, which answers "should we keep funding this". The second is the age of each article, which answers "which topics should we repeat". A post published nine months ago that appears in the form answers of three closed deals deserves a refresh and a sibling article. A post with traffic and no deals tells you the topic attracts readers who do not buy.

Cancelling the program resets this clock, which is why leaving a provider mid-year costs more than the invoice suggests. The consequences for traffic are laid out in what happens to your rankings if you fire your SEO agency.

Two people studying printed financial documents at a table
Reading deals one by one is slower than a dashboard and far harder to argue with

When should you distrust your own number?

Whenever the sample is small. Ten closed deals a quarter means one deal moves the ceiling by ten percent, and a single large contract can make a weak quarter look strong. Look at four quarters together before you change a budget.

A form answer of "your blog" often names the article read last, which may not be the one that persuaded. A report that quotes one precise ROI figure with no range hides its assumptions, so ask which window and which rule produced it.

Ownership of the finished articles also affects the math, because a blog you can keep paying off for years behaves differently from one that disappears with the contract. Questions about the posts your SEO agency wrote belong in the same conversation as the ROI one.

Sources

  1. Sotros Infotech, "How to Measure B2B Content Marketing ROI: The Multi-Touch Attribution Framework for 2026" (single-touch versus multi-touch share of pipeline, practitioner audit, 2026).
  2. 1827 Marketing, "Dark Social and Multi-Stakeholder B2B Attribution" (private sharing and buying committees, 2025).
  3. Worked examples in the ledger and the sales-cycle illustration are our own arithmetic and not client data.

Check your site in a minute

Paste your address: we'll show your score out of 100, whether an AI assistant names you, and which searches your site has no page for. Free.