← Blog · Content operations
Annual Content Refresh Is Dead. Your Blog Needs Governance.
Every article you published is a claim about your product, your market, and your position. Most of those claims expire quietly. A yearly cleanup was enough when the market moved yearly. It no longer does.
Your blog breaks in four ways
A blog is usually treated as an accumulation problem. Publish, index, rank, repeat. The archive grows and the assumption is that growth is the whole job. It isn’t. An archive is a liability that appreciates only if it is maintained. Every post you’ve ever published is live, crawlable, and quoted back at prospects as your current view of the world, regardless of when you wrote it. When we audit a client’s existing archive before taking over, the damage tends to fall into four categories. None of them are caused by bad writing. They are caused by time.Features land in one post, then vanish from the rest
A product ships something significant. Marketing writes the launch post, maybe a use-case article, and moves on. Meanwhile forty older articles describe workflows that the new feature has changed, shortened, or made unnecessary. A prospect reading an eighteen-month-old comparison piece sees a product that no longer exists. The launch post corrected the record in exactly one place. The rest of the blog still argues against it.Positioning moves, the archive doesn’t
Companies reposition more often than they admit. The ICP narrows, the category name changes, the “for teams” story becomes the “for enterprise” story. The homepage updates in a day. The blog keeps selling the old position for years, because nobody budgeted for rewriting two hundred intros. This is the most expensive form of decay, because it contradicts the thing sales is currently saying on calls. We wrote about why a blog needs one core theme; an ungoverned archive ends up with three, layered by era.Competitors catch up, and your differentiator becomes table stakes
The post from two years ago that led with “the only platform that does X” was true when it shipped. Then three competitors shipped X. The claim is now either false or, worse, it signals that you haven’t noticed. And the differentiator that replaced it exists only in the posts written after the change. Your most-read evergreen articles, the ones that actually rank, are the ones making the weakest case.Links and cross-references rot
Screenshots of old UI. Pricing pages that have moved. Integration partners that were acquired. Internal links pointing at posts that were quietly unpublished. Each one is small. Together they tell a reader that nobody is home.The common thread: a blog is written once but read continuously. The writing was a project. The reading never stops. Maintenance has to match the reading, not the writing.
Why the annual refresh made sense
The conventional answer to all of this is the annual content refresh. Once a year, usually tied to planning season, someone pulls a traffic report, picks the top thirty posts by sessions, and updates them. New date, new screenshots, a paragraph about the new feature, republish. This was a reasonable design for the market it was built in. Products shipped major releases a few times a year. Competitors published a handful of substantial articles a month. Positioning changed with funding rounds, not quarters. The rate of decay was slow enough that a yearly pass caught most of it before a prospect did. The refresh cadence was calibrated to the decay rate, and the decay rate was annual. It also had a quiet flaw that the slow market hid. The refresh selects by traffic, not by risk. The post that ranks first gets updated. The post on page two that a serious buyer finds during month four of their research, the one that says you don’t support SSO, stays wrong, because it never made the top-thirty list.What changed the clock
Two things accelerated at once, and both of them are downstream of AI. First, competitors ship faster. Feature velocity in B2B SaaS has visibly increased, and a differentiator now has a shorter life before it is matched. The window in which “only we do this” is true has compressed from years to quarters, sometimes months. Second, competitors publish faster. The cost of producing a plausible-looking article has collapsed. Whether that content is good is a separate question, and mostly it isn’t, but volume alone changes the landscape your archive sits in. A year-old post is now surrounded by hundreds of newer ones. Its age is more visible, and its stale claims are more easily contradicted. The refresh cadence did not move. The decay rate did. The gap between the two is where your blog is currently losing arguments you don’t know it’s having.The annual refresh was never a strategy. It was a cadence that happened to match the market. The market stopped matching.
What governance actually looks like
The alternative is not a faster refresh. Doing the same traffic-sorted cleanup every quarter instead of every year still selects the wrong posts and still reacts after the damage is done. What’s needed is closer to how a product team manages a codebase: every published article is an owned asset with known dependencies, and any change to the product, the positioning, or the market triggers a review of the articles that depend on it. The difference between the two models is clearest side by side. The refresh treats the archive as a list of URLs to be visited periodically. Governance treats it as a system with inputs, and a change to any input is an event that propagates through the archive.
Refresh versus governance: what triggers work, and on what
In practice, governance has three components, and all three are unglamorous.
| Dimension | Annual refresh | Continuous governance |
|---|---|---|
| Trigger | The calendar | A product release, a positioning change, a competitor move, a broken dependency |
| Selection | Top posts by traffic | Every post that depends on what changed, regardless of traffic |
| Unit of work | Rewrite the whole post | Fix the specific claim, screenshot, link, or section that is now wrong |
| Who knows what’s in the archive | Nobody, until the audit | A maintained map of which posts make which claims |
| Failure mode | Eleven months of drift between passes | Review backlog, if the map is neglected |
| Relationship to new content | Separate project, separate budget | Part of the same editorial operation |
- A claims map. For each published article, a short record of the product facts it asserts, the positioning it assumes, and the competitive claims it makes. Not a summary. A list of things that could become false. This is the thing nobody builds, and it is the whole mechanism. Without it, a product change cannot be traced to the articles it invalidates.
- Change intake. A standing agreement that product, sales, and leadership tell content when something moves. Release notes, positioning memos, win/loss notes where a competitor’s new feature came up. Content does not need to be in the room for the decision; it needs to hear about it the same week.
- Review as routine, not project. A small, recurring slot, weekly or fortnightly, in which the changes that came in are matched against the claims map and the affected posts are fixed. Most fixes are a sentence. Some are a section. Occasionally a post is retired. The point is that it happens continuously and nobody has to “find time” for it.
Who owns it
This is where most companies stall. Governance falls between roles. Product doesn’t own the blog. Marketing owns the calendar, not the archive. Freelancers own nothing past the invoice. The work is real but it has no natural home, so it doesn’t happen. It needs one owner with three things: long tenure with the product, so they recognise when a change matters; authority over the archive, so they can edit anything without a ticket; and a reason to care about posts that aren’t theirs. That last one rules out most freelance arrangements and most agencies structured around deliverables, because the economics reward new words, not corrected ones.
“We refresh our top content annually whenever the product, the positioning, or the market changes, and we know which posts that touches before it happens.”
The phrase “content governance” sounds like bureaucracy. It is the opposite. Bureaucracy is the annual audit: a large, dreaded, periodic event that everyone avoids until it’s forced. Governance is small, continuous, and boring, and it is the only thing that keeps a growing archive from turning against you.
Key takeaways
- An archive decays in four ways: features land in one post, positioning shifts, differentiators get matched, and dependencies rot. None of these are writing problems.
- The annual refresh was calibrated to an annual decay rate. AI has compressed feature and publishing cycles, and the decay rate is no longer annual.
- Governance is event-driven, not calendar-driven. A change in the product, the position, or the market triggers a review of the posts that depend on it, regardless of traffic.
- The mechanism is a claims map: a record of what each post asserts, so changes can be traced to the articles they invalidate.
- It needs one long-tenured owner with edit authority over the whole archive. Deliverable-based arrangements structurally cannot provide this.