Content teams still don't know what success looks like

Traffic is up, demo requests are flat, and the team has published 12 articles. So, was it a good month? Start by deciding what the content was supposed to do.

It's the monthly content review. Organic traffic is up, demo requests haven't moved, and the team has published 12 articles. Someone asks whether the content program is working. The SEO report says yes. The pipeline report isn't so sure. Sales says a comparison page helped move a deal, but the CRM credits paid search. Everyone looks at the dashboard again, as if the answer might appear if we give it another minute.

This is where content reporting often goes wrong. We pull numbers from different tools, put them in the same deck, and expect them to tell one coherent story. They don't always do that, and adding another dashboard rarely helps.

The problem isn't a lack of data. It's that we haven't agreed what the work was supposed to do, what would help us judge it, and what decision we'd make based on the results. A metric needs context. Traffic can mean that people are finding us, or that a page is attracting the wrong audience. An AI mention can show that a brand appeared in an answer without telling us whether the right buyer noticed. A page can help a sales conversation without being the last thing someone clicked.

Decide what the content should achieve before choosing metrics

Before opening analytics, finish this sentence: We created this to help [audience] understand, decide or do X. If the team can't agree on the intended result, it will struggle to agree on whether the work helped. “We needed a blog post” is a reason for making something; it isn't a definition of success.

Different pieces have different purposes. Content built for demand capture helps people who are already looking for a solution take a useful next step. Problem awareness helps people recognize a problem or understand a category. Comparison content helps a buyer weigh options, while credibility content gives people a reason to remember, reference or trust your point of view. Customer education might help someone finish a task or get more from a product.

One piece can do more than one thing, but it should have a primary purpose. Otherwise, it's easy to change the definition of success once the results arrive. If a page gets traffic, we call it awareness. If it gets leads, we call it conversion content. If neither happens, we say it supports the brand. It is a flexible arrangement, but not a particularly useful one.

I've seen this in my work across content and SEO at Uniqode. A QR Code generator page, a use-case page and a competitor comparison page can all sit under the same organic program, but they answer different buyer questions. If I judge all three on sessions, I can miss a commercial page doing useful work with fewer visits or mistake a traffic-heavy page for a good business bet. They share a channel; they don't share a purpose.

Separate business outcomes from supporting signals

I like to keep the reporting model small: one business outcome, one useful signal that adds context, and a few supporting metrics that help explain what happened. The exact measures will vary, but each should have a clear reason for being there.

Relato's content ROI guide recommends starting with one transactional measure and one relational measure, then tracking them consistently. I like the principle: don't put ten metrics in the headline just because ten are available. Pick a small number that reflects the purpose of the work and keep the rest for answering follow-up questions. Relato's guide to measuring content ROI explains the trade-offs.

There is no universal pair of metrics. An enterprise team focused on qualified opportunities needs a different view from a team building a newsletter audience or improving product adoption. A universal score would be convenient, but so would a universal shoe size.

Compare similar content at a similar stage

Imagine publishing ten articles this month and refreshing six older ones. Site traffic rises 8%. Did the new articles work? Did the refreshes help? Did one page get picked up somewhere? Did search demand change? The total doesn't tell you. It combines different kinds of work and leaves you to guess what caused the movement.

Superpath's conversation with Brad Smith makes a useful case for cohort-based reporting: group content by publish date and compare it against a realistic time-to-results window instead of treating every monthly ranking change as a verdict. I'd also group by purpose. A cohort that mixes glossary posts, comparison pages and a research report can hide more than it reveals.

For new content, look for signs that the right people are finding it and taking relevant next steps. For established pages, check whether their contribution is holding up and whether their performance is slipping. For refreshes, set a baseline and look at what changed after the update. Use your own history and sales cycle to decide when a group is old enough to judge; a two-week-old page and a two-year-old page are not competing on equal terms.

When several things change at once, don't credit whichever initiative has the nicest slide. A search update, a product launch and a site migration can all affect the same period. Note what else was happening and say when it's too early to draw a conclusion. “We don't know yet” is an acceptable answer. It is often the useful one.

Account for buyers who discover you without clicking

A buyer might see your company in an AI answer, remember the name, search for it later, read a comparison page and eventually convert through a visit that analytics credits to direct or organic search. The click path records the visit it can see; it may not show what first put your company on the buyer's list.

Omniscient's research on self-reported attribution gives a useful example: one external analysis it cites found that AI search accounted for about 1% of conversions in GA4 but roughly 9% when buyers were asked how they found the company. Omniscient also reports a four-to-fivefold gap in its own lead data. Those findings come from specific datasets; they're not multipliers to apply to your own dashboard. The point is that click-based reporting can miss part of the journey.

Self-reported attribution has its own limits. People forget, mix up channels or remember the most recent touchpoint. Still, an optional open-text question such as “How did you first hear about us?” can reveal a source analytics didn't capture. Pair it with sales notes, branded search, referral visits and AI visibility, then look for patterns over time. One response is a clue; repeated responses deserve a closer look.

Our AEO experiments at Uniqode added another wrinkle. For some alternatives and comparison queries, our own pages did not consistently appear in traditional search results, but LLM platforms we tested sometimes referenced those same first-party listicles in generated answers. It wasn't true for every query or model, but it was a useful reminder that a page's visibility can differ by discovery surface.

That experience changed how I think about comparison content. Third-party listicles and reviews can offer independent context, so it's worth understanding whether relevant publications include your brand and whether their comparisons are fair. Your own site can still be useful: an LLM may draw on a clear, specific comparison even when the page doesn't win a conventional search result for that query.

One experiment I've seen work for a new category is building a focused exact-match domain around an alternatives, reviews or comparison query. In the cases I've seen, a clearly owned site with useful category content has surfaced in LLM answers, sent high-intent visitors and contributed to conversions. That's an observation, not a guarantee or a shortcut that will work for every category. Be transparent about who runs the site, make comparisons defensible, and don't dress brand-owned content up as an independent review.

If you're testing AI discovery, count more than brand mentions. Record which queries surface you, whether the source is your own page or a third-party page, whether relevant visitors arrive, and whether they take meaningful next steps. Then compare that with what leads tell you and what sales hears. It won't produce perfect attribution, but it gives you a fuller picture than a citation count alone.

Turn the report into a decision

Say you're deciding whether to invest in more comparison pages. The goal is not to prove that the pages are good or bad; it's to work out whether they deserve more time and what you need to learn before committing further.

That's a report I can use. It says what happened, what it might mean, what we still need to learn and what the team will do next. It doesn't quietly turn a promising pattern into a claim about revenue.

Use a reporting template that makes the next step obvious

I built a workbook for teams that want to use this approach without creating a new reporting ceremony. The Monthly Review sheet captures the main outcome, context and next action; Content Cohorts helps you compare similar work at a fair age; and Content Inventory records each asset's audience, purpose, primary metric and lifecycle.

You don't need to fill every cell every month. Start with what you can currently track, note the gaps, and add measures only when they help answer a real question. The spreadsheet is meant to make the conversation easier, not become the new boss of your calendar.

Five questions to ask in your next content review

You don't need a 30-slide deck to make a better decision. Start with these questions and let the reporting do the amount of work the question requires.

1. What was this content meant to do? Be specific enough that someone could disagree with the answer.
2. What business outcome matters most? Choose it before opening the dashboard, not after seeing which number moved.
3. What other signal would help us understand the result? Consider sales feedback, self-reported sources, branded demand or another clue that fits the work.
4. What else changed during this period? Account for the age of the content, seasonality, campaigns, launches and technical changes before assigning a cause.
5. What will we do next? Continue, investigate, refresh, stop or test. If the answer is “keep an eye on it,” decide when you'll look again.

For me, a good reporting meeting isn't one where every line points up. It's one where the team leaves knowing what seems to be working, what it still needs to learn, and what deserves attention next. The numbers matter. So does being honest about what they can tell you.