A SaaS blog that compounds instead of decaying

Take your blog’s URL list, sort by organic sessions over the last twelve months, and find the point where the cumulative total hits 90%. On every B2B site I have done this for, that point falls somewhere between the 8th and the 30th post.

Everything after it is maintenance cost. Some of it is worse than neutral, because it competes with the pages you actually care about.

The publishing habit problem

Most SaaS blogs exist because someone decided the company should publish weekly. The cadence became the goal, the calendar became the strategy, and now there are 240 posts and a person whose job is to keep the streak alive.

Nothing in that process asks what any individual post is for. So the posts are all vaguely about the category, they overlap heavily, they link to each other in a soup, and the four that matter are buried.

The fix is not to publish less as a general policy. It is to require, for every piece, an answer to a specific question before it gets commissioned.

The job field

Every brief we write has a field called “job.” It has to be filled with one of a small number of things:

Rank for a specific commercial cluster and convert. Earn links, with a named target publication type. Add topical depth to a cluster where we are ranking fifth and need weight. Answer a question sales gets on every call, so the AE can send it. Establish a position that differentiates us in a category where everyone sounds the same.

If none of those fit, the piece does not get written. “Because it is an interesting topic” is not a job. Neither is “we have not published in two weeks.”

This single field cut one client’s output from eleven pieces a month to five, and organic demo requests went up 40% over the following two quarters. Some of that was the additional refresh capacity the cut freed up.

Internal linking, which does more than people think

A blog is not a list of posts. It is a graph, and the shape of the graph decides where authority accumulates.

The default shape is bad: chronological listing pages, a “related posts” widget picking by tag, and whatever links the writer happened to add. Authority spreads evenly across everything, which means it concentrates nowhere.

What works is deliberate hierarchy. Two or three substantial hub pages per major topic, each linking down to supporting articles. Every supporting article linking back up to its hub and across to one or two siblings. And every cluster linking into the commercial page it exists to support, with descriptive anchor text.

The last part is the one people skip. An article about keyword research that never links to your keyword research product page is doing half a job.

We audit this with a simple crawl: for each commercial page, how many internal links does it receive, from where, and with what anchors. On most sites the homepage and the blog index have hundreds of internal links and the pages that produce revenue have four.

The quarterly prune

Every three months, every URL gets sorted into one of four buckets.

Keep and leave alone: performing, current, no overlap. Usually 20% of an inherited blog.

Refresh: was performing, now decaying, or has stale data. This is where the fastest returns live. A refresh means re-checking every claim, updating screenshots and numbers, adding what has changed, improving the internal links, and republishing with an updated date. It costs maybe a third of what a new piece costs and on decaying pages it recovers traffic much faster.

Merge: two to four thin posts on adjacent topics that should be one substantial page, with the others redirected. The most emotionally difficult category and often the most valuable.

Remove: no traffic, no links, no strategic purpose, and no realistic path to any of them. Redirect to the most relevant surviving page, or let it 410 if nothing fits.

On a 200-post inherited blog the typical split is 40 keep, 50 refresh, 60 merge, 50 remove. Companies find the remove number alarming. The traffic loss from deleting pages that get under ten sessions a year is a rounding error, and the crawl efficiency and topical clarity gains are real.

Dates, and whether to show them

Show them. Hiding publication dates to avoid looking stale is a trick readers see through, and in B2B the reader specifically wants to know whether the pricing comparison they are reading is from this year.

What you should do instead is earn a recent date by actually updating. Show “Updated March 2026” and mean it. Changing the date without changing the content is the version that erodes trust when someone notices the same typo from 2023.

What a good blog looks like at steady state

Forty to eighty live posts, not four hundred. Four to eight new pieces a month, each with a stated job. Six to ten refreshes a month. A clear hub structure where you can name the three or four topics you are trying to own. And an internal link graph that pushes authority toward the pages that make money.

Also, ideally, an editor willing to say no. That role is worth more than an extra writer and almost nobody staffs it.