How to Tell Which Tech Publications Deliver the Most Readers per Mention

Readership per mention is the average number of readers one article delivers at a given publication: the total publisher-reported readers for that outlet's articles in a given theme, divided by the number of those articles with a non-null readership value. It measures what a single placement is worth in readers, not how large an outlet's audience could be. Ranked that way, outlet size and outlet efficiency can come apart.

This page does not publish a ranking of tech outlets. It shows how to build one from your own coverage, why the usual reach-based lists point you at the wrong outlets, and which checks keep the result honest.

Disclosure: Delve (delve.news) publishes this page and provides the article-level publisher readership the method uses. The arithmetic works with any article-level readership source.

For why the distinction matters in coverage quality and AVE replacement, see coverage quality benchmarks and AVE is dead.

What you need to know — TL;DR

Which tech publications deliver the most readers per mention?

It depends on your theme and window, and it is not reliably the outlets with the biggest brand or the largest monthly audience. The only way to answer it is to divide the readers an outlet's articles actually drew by the number of articles that carry a readership figure.

How is a mention counted?

One article counts as one mention, however many times a brand appears inside it. The average divides by articles with a non-null readership value, so a genuine zero stays in the denominator.

Why not rank by monthly audience?

Monthly audience estimates how many people could see something from an outlet. Readership per mention measures how many people read the individual articles in the set.

What should change in pitching?

Rank outlets by mean readership inside your own theme, with a floor on sample size, rather than by masthead prestige or estimated monthly audience.

Where reach rankings mislead

Lists of the biggest tech publications sort by estimated monthly traffic, search reach, or editorial reputation. All three describe the outlet. None describes what one placement in it delivers.

Two outlets can look the same on reach and behave very differently per article. A site that publishes into your theme every day spreads its readers across hundreds of pieces. A site that publishes a few pieces a week on the same theme may send far more readers to each one. A reach ranking cannot see that gap, because it never looks at the article.

Impressions make the error worse. Impressions answer how many people could have seen something from an outlet. Publisher readership answers how many people read this article. AVE and impression logic treat those as interchangeable, which is why AVE is dead puts readership-weighted quality in the replacement stack.

Key takeaway: reach describes the outlet. Readership per mention describes the placement. Only the second tells you what one story was worth.

How to calculate it

1. Fix the universe. Choose one: all articles tagged to a theme, or a locked list of outlets plus a theme filter. State which one in the report.
2. Fix the window. Use a stated window, for example the trailing 12 months to a named date, and keep it for every re-run.
3. Count one mention per article. An article counts once however many times a brand appears inside it. In a theme ranking the article is in the set because it matched the theme filter, not because it named you.
4. Pull article-level readership from publisher analytics. Do not fill gaps with estimated reach.
5. Divide correctly. Readership per mention equals total readership divided by the count of articles with a non-null readership value at that outlet. Publish both counts: articles tracked, and articles with readership.
6. Apply a minimum sample. Drop outlets below a stated floor, for example 50 articles with a non-null readership value in the window. Set it where your own data stops being noisy.
7. Keep genuine zeros, drop nulls. A null means no data and leaves the calculation. A true zero is a real low value, and removing it lifts every average.

For a brand-scoped list instead of a theme ranking, run the same arithmetic filtered to articles that name your brand. The denominator changes and nothing else does. Keep the two cuts separate, because an outlet can be efficient in a theme and thin on your coverage specifically.

Key takeaway: publish both counts. The average divides by articles carrying a readership value, and anyone checking your work by hand will use whichever denominator you printed.

A worked example (illustrative numbers)

The outlets and figures below are invented to show the arithmetic. They are not measured data.

Three things show up in this table.

• Scale and efficiency separate. Outlet A has the largest total at 18 million readers. Outlet B has about half that total on a fifth of the articles, and 2.5 times the readers per article.
• The denominator matters. Dividing Outlet A's total by articles tracked gives 18,000, not 20,000. Dividing by articles with readership gives 20,000. A reader who checks your page by hand will use whichever number you print, so print both counts.
• The minimum sample protects the ranking. Outlet C posts the highest average, on 12 articles. One breakout story can produce that. It is excluded until it clears your minimum, 50 in this example.

What this means for pitching priorities

1. Build a theme-specific outlet list sorted by mean readership per mention, with a minimum article count so a single viral outlier cannot crown a random site.
2. Keep a second sort on total readership for launches and category moments where scale is the point.
3. Stop treating a high-volume outlet as a readership win by default. Check the article's own readership once it appears, or the outlet's trailing average in your theme.
4. Use efficiency for prioritization, not vanity. A trade outlet with a lower absolute average can beat a national brand on relevant readers, so pair this ranking with audience fit.
5. Re-run by theme. AI, cybersecurity, and consumer tech will not share an efficiency order.

Key takeaway: sort by efficiency to choose targets, by total readership when scale is the point, and never by masthead alone.

Limits of the method

• Coverage is partial by design. Only outlets with article-level publisher analytics carry readership. Outlets without it are absent from the ranking rather than ranked low, and estimated reach is never substituted to fill the gap.
• Theme tagging is often model-assigned. Technology, AI, and Innovation overlap, and some articles carry several themes. State whether an article counts once per theme or is deduplicated across themes.
• Mean is not median. A few breakout stories lift an outlet's average, and a per-article mean can hide the articles that outperformed it by a wide margin. Report the median beside the mean, or winsorize the top one percent.
• Not causal. A high average does not mean a pitch caused the traffic. It means placements at that outlet, in this theme, historically drew more readers per piece.
• Not message quality. An outlet can deliver readers for an off-message or negative story. Pair this with key-message pull-through and sentiment when the question is campaign success rather than outlet efficiency.
• Windows drift. Averages move with the news cycle. Fix the window and keep it for the next run.

How Delve measures readership

Delve reports article-level unique visitors sourced from publishers, which is the number behind readership in this method and in the product. Readership data is included on the Starter plan. Reach data is a separate feature on the pricing page and answers a different question: the audience a placement could have reached, rather than the readers a specific article drew. The two are not interchangeable, and this method uses only the first.