How to Score a Blog Topic Before You Spend a Draft on It

You have budget for three drafts this month. Your research tool just handed you thirty "high-opportunity" topics. Most of those topics will burn a draft and rank nowhere, because the tool scored them against its own criteria, not yours.

The fix is not a better tool. It's a fixed score you apply before anything gets drafted: three checks, in the same order, on every topic, across every client site. This post gives you the checklist. By the end you'll be able to score any topic idea in a few minutes and know exactly why it passed or failed.

Why topic research tools disagree — and why you need your own score

Run the same keyword query through two research tools and you'll get two different lists of "top opportunities." That's not a bug. Each tool optimizes for something different: one weights volume, one weights difficulty, one weights commercial intent. The output reflects the tool's bias.

Even practitioners can't agree on what matters. Experiments In Search lists four keyword-selection criteria — search volume, ranking difficulty, search intent alignment, and business relevance (experimentsinsearch.com). AEO Autopilot uses three tests instead: buyer search demand, realistic ranking chance, and mapping to something you actually sell. Those overlap, but neither list is complete, and neither is yours.

The takeaway: treat tool output as raw candidates, never as verdicts. A tool that hands you thirty topics has done the gathering. The judgment — does this earn a draft — is still yours, and it needs to be systematic. Hence the three scores below.

Score 1: Source freshness — is the evidence recent enough to build on?

Before you care about anything else, check whether the topic's supporting facts come from sources that are current and verifiable. A topic with high search volume built on recycled facts from years-old listicles is a trap: you'll draft it, fact-check it, and find the evidence is soft.

The practical check takes ten minutes. Open the top three to five sources your research tool cites for the topic. Look for three things: a date, substantive content behind the headline, and independence. If the sources are undated, thin, or all citing each other in a loop, the topic is stale no matter what the volume says.

The pitfall here is trusting a tool's confidence score as a freshness signal. Most tools don't date their evidence, and a confidence number tells you nothing about when the underlying sources were published. You have to open the sources yourself.

Why this comes first: a draft built on stale facts fails fact-check later, and a failed fact-check means a rewrite. That's a full draft slot spent on a post that can't ship on schedule. Checking freshness before drafting costs minutes; catching it after drafting costs a draft.

Score 2: Audience fit — does this topic serve this site's reader?

Audience fit is the score most agencies skip, because doing it properly requires knowing each client's reader, which usually means a written brief nobody maintains. So topics get scored against a generic "content marketing reader" who doesn't exist, and posts land that are fine in the abstract and useless for the actual audience.

The fix is a per-site audience profile built from the site's own pages. What does the site sell? Who does it address? What tone does it use? You can pull all of that off the existing content without asking the client for anything. Then score each topic against that profile, not against the industry in general.

The pitfall: reusing one client's topic list on another site in the same niche. Same industry, different audience, different fit. A "best payroll software for startups" post might be a strong fit for a fintech blog aimed at founders and a dead miss for a payroll provider whose readers are HR managers at mid-size companies. The niche is identical. The fit is not.

Keep the profile editable. When a client pivots, or when the profile starts reading wrong, fix the profile before scoring any new topics. A stale profile silently poisons every fit score downstream.

Score 3: Search intent — is anyone actually looking, and can you rank?

The third score has four parts, and they're quick.

First, intent alignment. Does the searcher want what this post answers? Experiments In Search lists search intent alignment alongside volume and difficulty for a reason: the criterion is about matching what the searcher actually expects to find. If the searcher wants a tool and you're offering an explainer, the topic fails.

Second, volume, as a floor rather than a goal. AEO Autopilot notes that 100 to 1,000 monthly searches is often enough for a small business blog, especially when searchers are close to purchase. Don't discard a topic because the number looks small; discard it because the number is zero or the intent is wrong.

Third, ranking chance, checked honestly. AEO Autopilot recommends targeting keywords where at least two or three page-one results come from sites of similar size to yours. If all ten results are major publications, your post is competing for a spot that doesn't exist. Be strict about "similar size." A 40-page client site is not competing with a domain that has 40,000.

Fourth, specificity. AEO Autopilot's rule: if you can't write a complete, accurate answer in 50 words, the topic is too broad. Try it. If the 50-word answer turns into "it depends" three times, split the topic into two narrower ones or drop it.

The pitfall for the whole score: chasing volume on broad head terms where you can't compete. It's the most common failure in topic selection, and it's invisible until the post sits on page four for six months.

Running the checklist across every client site without it drifting

A checklist only works if it's the same checklist every time, for every site, run by anyone on the team. Three things make that hold.

Write the three scores as a fixed list: same questions, same order, every topic. Freshness, fit, intent. If the order varies by who's scoring, the results will vary too.

Score in batches. Run all candidate topics through the checklist in one sitting per client, then pick the few that clear the bar. Scoring one topic at a time as ideas occur invites inconsistency; batching makes the comparisons real, because you're choosing the best of a known set.

Watch for the urgency override. A client calls, a post is due, and a topic that would fail the freshness check gets waved through. Don't. A topic that fails freshness or fit this month will still fail next month; the only thing urgency changes is who gets blamed for the weak post.

Keep per-site scoring records. Note what you scored, what passed, what you drafted. Comparing your predictions against what actually ranked is the only way the checklist improves; otherwise you're running the same judgments forever with no feedback loop.

How to check the checklist is working

Track two numbers per site: topics scored versus topics drafted, and drafted posts that hit their target within 90 days.

Read the ratio between them. If your draft rate is high but your hit rate is low, the scoring is too loose; tighten the intent or fit thresholds until weaker topics stop passing. If almost nothing passes, the checklist is too strict, or the audience profile is off. Audit the profile first; a wrong profile fails good topics for the wrong reasons.

Review the scores monthly. Search behavior shifts, clients pivot, and a checklist you never revisit stops matching either one. The monthly review is where you adjust thresholds based on the hit-rate data, not on gut feel.

Where the pipeline does this for you

Everything above is doable by hand, and if you run two sites, do it by hand. If you run eight client sites, the audience-fit score in particular becomes the bottleneck, because it depends on a per-site profile you have to keep current.

This is where ContentRails fits. It researches topics by drafting ten ideas and having a separate critic score them (contentrails.ai), and it scores audience fit against a Project DNA profile built from each site's own pages — so the fit half of the checklist is measured against the actual site, not guessed from a brief nobody wrote. If the profile reads wrong, you edit it (contentrails.ai), same as the manual checklist says.

Projects stay separate, each with its own topics and posts, so one client's scores never bleed into another's. The pipeline itself is fixed and logged (plan, research, cited fact sheet, outline, draft, checks, editor), and every figure in a draft traces to a source the system actually read, which is the freshness check enforced at the draft stage rather than at scoring time.

Autopilot levels are set per project: Suggest topics for review-heavy clients, Full autopilot elsewhere, with you responsible for review before publication either way (contentrails.ai). The free plan runs one project end to end, three posts a month, no card required, so you can test the scoring loop on a single site before rolling it out across the roster.

The decision rule

A topic that fails any one of the three scores goes back to the idea pile, no matter how good the tool says it is. Stale sources, wrong audience, or unwinnable intent: one failure is enough. The tool's verdict is a suggestion; your checklist is the gate.

Next step: take your next ten topic ideas and run all three scores before drafting any of them. If you want the scoring done the same way every time, set a ContentRails project's autopilot to Suggest topics. Ideas arrive pre-scored, and the draft decision stays yours.