Every mid-market board I have sat in on this year eventually gets to the same question: are we on the right content stack for what is coming next. The ones that have money to spend hear it as a procurement decision. The ones that have already spent it hear it as a sunk-cost conversation. Both camps are nervous, and both are right to be. The stack you picked in 2023 was sold on headless and composability, and the market has quietly moved on from the vocabulary before most of those projects have shipped a second phase.
The reason I keep coming back to WordPress VIP in that conversation is not because I am loyal to the platform. I am loyal to a different thing, which is boring continuity. The stack that still looks sensible in three years is the stack whose vendor and whose analysts are telling the same story about where things go next, and where the governance and permissions already exist to accommodate it. For a buyer, that is almost always worth more than any one architecture label.
So I went looking at what the analyst who most buyers quote on CMS is actually saying this year, cleaned of vendor framing. Chuck Gahun is a VP, Principal Analyst at Forrester, and he wrote the Forrester Wave: Content Management Systems, Q1 2025. Below is what he has published, what it points at for a buyer, and what we would build underneath it on a WordPress VIP estate.
The money is in time to market, not in the architecture debate
Gahun’s framing of the Q1 2025 Wave, published on his Forrester blog on 24 March 2025, is the sentence most of my clients should have on a sticky note: time to market is now the primary growth driver, and the next generation of content management systems is being judged on how fast content and variants ship, not on which architecture diagram they fit into. He is explicit that headless and composable architecture have become building blocks for the AI-driven evolution rather than the destination.
That is a quiet but important shift. For the better part of five years, the enterprise CMS debate has been MACH-flavoured: pure-headless APIs, pulled-apart stacks, best-of-breed, bring-your-own-frontend. A lot of money went into those projects, and a lot of it is still sitting in half-shipped states. Gahun is not saying any of that was wrong, he is saying the question has moved on. If the actual growth driver is time to publish, every assembly step a composable stack adds between an editor and a live page is a cost. That cost used to be defensible when the payoff was flexibility. It is harder to defend when the payoff is also flexibility, after three quarters of integration work.
Time to market is now the primary growth driver.
Chuck Gahun, VP, Principal Analyst, Forrester, 24 March 2025
AI is already in production, and most leaders are not ready for it
The second thing worth sizing is the gap between where the technology has reached and where the organisation is. Shane Schick, writing on the WordPress VIP blog on 22 April 2026, pulls two figures that fit together uncomfortably: Deloitte’s research puts around 40% of AI projects in production, expected to double, while McKinsey’s survey work finds 86% of C-suite leaders feel unprepared for AI adoption. Those two numbers are not contradicting each other. They are describing the same organisation. The tooling has shipped faster than the people deciding how to use it.
What this looks like on the ground is familiar. A content lead has three AI copy tools her team already uses, none of them governed centrally. A product owner has shipped an agent prototype against the e-commerce catalogue, which nobody has reviewed for brand compliance. Legal is not sure what the retention policy is on the model’s inputs. The platform team is being asked, usually in the same week, to connect three different large language models and to pretend that is a strategy. The reason buyers are gravitating toward platforms with explicit governance primitives, rather than hobbled-together integrations, is not fashion. It is the only exit from this mess that does not require asking the board for another sixty thousand euro of unplanned headcount.
Composability was the building block, not the destination
The lever Gahun names most clearly in his Wave write-up is the collapse of the distance between core authoring and AI capability. Visual editors, personalisation controls and collaboration tools, which the composable era pushed out to the edges of the stack, are now centrepiece CMS features again. The reason is operational: an editor pasting content between four tools cannot ship fast enough to feed a market where agents also want to read the output and summarise it to a buyer.
WordPress VIP’s placement in the Q1 2025 Wave reads better once that lens is applied. Forrester marked it as a Strong Performer, with specific call-outs for its content compliance capabilities including FedRAMP certification, a solid roadmap across intelligence and platform enhancements, and the ecosystem of third-party plugins and add-ons. VIP’s own commentary on the Wave is here, dated 28 March 2025. For a buyer, the practical takeaway is less about the Wave quadrant and more about what Forrester graded WordPress on at all: compliance, roadmap and ecosystem. Those are three dimensions you cannot bolt onto a composable stack after the fact without rebuilding the governance you took apart.
Gahun has made the same argument in a less corporate register. In a conversation with Greg Kihlström on The Agile Brand in 2026, he said the business case is a traditional content strategy case, in what is more of a white space frontier. His position in that interview is that organisations should lean on existing tools and invest in content strategy rather than in a new technology implementation. If you are deciding where the next pound of budget goes, that is the recommendation the research covers.
The agent as audience is the lever nobody has fully cracked
The harder shift, and the one Gahun has gone further on since the Wave shipped, is in a February 2026 Forrester blog where he wrote that machines are now simultaneously content creators and, in deciding what buyers see, a primary content consumption audience. That is the sentence I would print and put on the wall of any marketing team still planning for a human-only reader. It also sets a bar most organisations will not meet by Christmas.
Honest about what does not work yet: there is no clean measurement layer for the agent-as-audience problem. Jodi Cerretani, CMO of WordPress VIP, said in the June 2025 webinar she ran with Gahun and Michael Khalili, Director of Product Marketing at WordPress VIP, that the tools that measure visibility today, hers included, mostly run on JavaScript in the browser, and the crawlers an agent is built on do not execute JavaScript. That is not a WordPress VIP problem, it is an industry problem, and the vendors pretending otherwise are papering over it. If you are evaluating a platform this quarter and the demo pretends to measure agent citation reliably, that is a prompt to ask a harder question.
What does work today is the plumbing: structured content, machine-readable fields, a canonical place a model can learn an entity from, and a governance layer that enforces consistency across the surfaces a brand does not own. That is the stuff you build before the measurement arrives, because the measurement will arrive against whatever you built.
What this looks like inside a WordPress VIP estate
Here is where we leave the analyst view and translate it to a ticket a platform team can book against. For an enterprise content estate the shape of a mid-market publisher or a multi-brand commerce group, the build we would scope on WordPress VIP has four layers, in the order we would ship them.
One, the content model. Agentic readers want fields, not free text. For a 2,000 to 10,000 SKU catalogue, or a newsroom with 60 to 150 taxonomies, the first pass is a content audit against the entities that matter: brand, product, author, event, place. Model each as a block or custom post type, not a WYSIWYG section. On VIP, we would use the built-in Parse.ly entity graph, the standard structured-data output, and the editorial workflow controls to make this the default state rather than the exception. Expect two to three weeks of senior editorial and platform time against a real catalogue, longer if the old estate has grown unmanaged for more than three years.
Two, the governance layer. Role-based access on AI capability is not optional any more. VIP’s platform now ships explicit permission scaffolding for the official OpenAI, Claude and Gemini provider plugins, which is what Shane Schick’s piece was pointing at in April. Centralising the key management, the model selection and the audit log at the platform level, rather than at the plugin or user level, is the difference between a buyer-safe estate and a legal exposure. Budget one to two sprints for this on a VIP estate; longer on self-managed hosting where you are also writing the retention policy from scratch.
Three, the publish pipeline. If time to market is the growth driver, the number to tune is how long it takes a signed-off asset to appear on a live page in every market it should appear in. On a VIP multisite with the usual compliance review steps, we have seen that go from 48 hours to under 20 minutes with a deployment pipeline that uses the VIP review gate for code and lets editors ship content directly through the standard editorial flow. The gain is architectural, not clever: fewer human handoffs, each one on a system designed to accept them.
Four, the measurement baseline. Even without a clean agent-citation metric, you can instrument today’s answer: how often does the brand appear in AI Overviews for queries you care about, what is the share of referral from assistant-driven search surfaces inside GA4, and what is the citation pattern across the big three assistants. We treat these as a baseline against which to measure the first three layers, not as a KPI anyone should hit yet. Four to six weeks of instrumentation plus a quarter of observation before anyone draws conclusions.
The whole build on an estate of that size runs eight to twelve weeks of senior time, assuming the content is in reasonable shape to start with. Longer if the taxonomy is a mess, which it usually is. The honest reason this is cheaper on WordPress VIP than on a torn-apart composable stack is not that the platform is magic. It is that the compliance, the governance and the editorial workflow are already there, in the same product, under one procurement contract, with one support channel. Forrester’s commissioned Total Economic Impact study puts the composite three-year return at 415%, with $2.8 million net present value and a 44% lower total cost of ownership against the proprietary platforms it tested. We treat vendor-commissioned TEIs with the usual caution, but the direction is in line with what we see on actual migrations.
What is still unresolved
Gahun is more cautious in the podcast interviews than in the Wave write-up, and I would stay cautious with him. The agentic shift is real, the measurement is immature, and most of what has been written about buying for it has been written by vendors who have something to sell on the other side of the recommendation. If you are a mid-market or enterprise buyer thinking about the next content stack cycle, the useful question is not whether to go headless or composable or hybrid. It is whether the platform you are buying has governance, compliance and workflow primitives that will survive the next two product cycles without a rip-and-replace. On that test WordPress VIP is a serious answer, and the reason is boring, which is usually a good sign.
If you are evaluating a move onto WordPress VIP, or trying to decide whether your current estate can get there without a rebuild, we are the people who do that work. Start with the WordPress VIP agency page; if the shape of the problem fits, there is a short route to a conversation from there.
Last modified: October 19, 2026
