Something breaks on a Tuesday. The form stops sending, or the site slows to a crawl, or a page returns an error nobody recognizes. The first question in the room is who to call, and the honest answer is a firm that built the site four years ago, a developer who left in 2023, and a hosting account in the name of somebody that nobody has signed into since launch.
The site is not badly built. It simply has no owner, and an unowned platform degrades quietly for years and then fails all at once, usually on a deadline.
Where this sits. This is the sixth and last of six posts on digital strategy. It is the area associations think of first, and the only one on the list that is purely infrastructure.
Getting the Proportion Right
Two opposite mistakes are common and they cost about the same.
The first mistake. The platform absorbs the entire digital conversation. Every strategy meeting becomes a conversation about the CMS, and the organization rebuilds every four years, each time expecting the rebuild to answer questions that were never about the platform. Member data, measurement, discovery and revenue are untouched by a rebuild, which is why the second one disappoints.
The second mistake. The platform is treated as the problem of somebody else until it produces a bill or a breach. Hosting renews automatically, updates are applied by whoever notices, and the first honest assessment of the platform happens in an incident.
The proportion that works. The right proportion is boring. This area should take a predictable, modest amount of money and attention every month, permanently, and should almost never be a strategic topic. When it becomes a strategic topic, it is usually because it has been neglected long enough to force one.
Four Questions That Define a Platform
Strip away the vendor language and a platform assessment reduces to four questions.
- Is it stable? Does it stay up, and does it stay up under the load of the one day a year that matters, which for most associations is the morning registration opens for the annual meeting.
- Is it secure? Not whether anyone has attacked it. Whether updates are applied on a schedule by a named party, whether anyone would notice a compromise, and whether you could restore from a backup that has actually been tested.
- Is it maintainable? Can a competent developer who has never seen it make a change without archaeology. Sites fail this quietly, through years of small customizations by different hands, and the symptom is that every quote you get is higher than you expected.
- Is it owned? Is there a named person or firm accountable for it, with access, with a budget line, and with an obligation to tell you when something needs doing. This is the one most associations fail.
End of Life Is a Calendar, Not a Surprise
The most avoidable platform emergencies are the ones published years in advance. Two are in front of associations right now, and both have documented dates.
- Drupal 10. The Drupal project states that Drupal 10 reaches end of life on December 9, 2026. After that date it stops receiving security coverage from the project, which for an association running Drupal 10 is a decision that has to be made this budget year rather than next.
- PHP 8.2. The PHP project lists security support for PHP 8.2 as ending on December 31, 2026. Active support for it ended two years earlier. Hosting platforms generally stop offering versions after that point, on their own schedule and not yours.
The point of naming them. Neither of these is news, and neither is negotiable, and both will produce a scramble at organizations where nobody is responsible for reading release calendars. That responsibility is the whole of what this area asks for most of the time.
What Gets Blamed on the Platform and Is Not
Because the platform is the visible layer, it absorbs blame for failures that originate somewhere else, and that misattribution is expensive because it produces rebuilds that fix nothing.
- Nobody can find anything. Almost always a content and labeling problem, sometimes a search configuration problem, and very rarely a CMS problem. Replacing the CMS moves the same content into a new template.
- The site is slow. Usually third-party scripts, oversized images and unoptimized media rather than the platform itself. The platform is often the only part of the page that is fast.
- Staff cannot update anything. Frequently a governance and training problem. A capable CMS that nobody was taught to use, with no agreement about who may publish what, behaves exactly like a bad CMS.
- We cannot integrate with the AMS. A real platform problem some of the time, and a permissions, licensing or vendor-contract problem the rest of the time. Worth establishing which before scoping anything.
How to tell the difference. The diagnostic question is simple and unpopular: if we rebuilt this on a different platform tomorrow, would this specific complaint still be true? For most of the complaints on that list the answer is yes.
What Good Looks Like
The platform area is in reasonable shape when these are true, and unlike the other five areas, all of them are achievable inside one budget cycle:
- You can produce a one-page inventory. Hosting, domains, certificates, the CMS, the major plugins or modules, the version of each, and the renewal date of each. One page. Most organizations do not have it and it takes an afternoon.
- Your domains and hosting are in your name. Your organization, not your agency, not a former employee, and not an email address that forwards to somebody who left.
- Updates happen on a schedule. With somebody named, on a stated cadence, and a record of what was applied and when.
- A restore has been tested in the last twelve months. A backup that has never been restored is a belief rather than a backup. Test one this year.
- Somebody reads the end-of-life calendars. Not the day it becomes urgent. The board should hear about Drupal 10 and PHP 8.2 style deadlines in the budget cycle before they land.
How to Tell Where You Stand
Four questions, and every one of them is answerable today:
- Who do you call? Not in an emergency. On an ordinary Tuesday, when something needs changing.
- Who can log into your hosting account? Log in. If nobody can, that is not a small problem, and it is a problem that gets much larger during a vendor transition.
- What version of the CMS are you on and when does it stop being supported? If the answer is a number rather than a date, you are managing this by hope.
- What is the monthly cost of keeping this running? If your platform spending is invisible until it is an invoice, it will keep surprising you.
The Boring Inventory, and Then the Whole Series
Start with the one page. Hosting provider and who holds the account, domain registrar and expiry, certificate and renewal, CMS and version, the plugins or modules you would not survive losing, the backup location and when it was last restored, and a name against each line. Nothing on that list requires a vendor and the exercise usually surfaces at least one thing nobody knew, most often a domain or a certificate renewing into an account nobody can reach.
That closes the series. Six areas: member data, measurement, discovery, member experience, revenue beyond dues, and the platform underneath. The reason we work through all six rather than the one somebody asked about is that the weakest one is rarely the one in the request, and fixing the requested area while the weakest one goes untouched is how organizations end up with a good website and the same problems. If you want the version of this assessment we run, it takes about two weeks, it covers the six areas in order, and it ends with a short document naming where you actually are rather than what we would like to sell. It is included in our partnership plans because it has to be repeated as the systems and the plan change, and a one-time assessment goes stale faster than anyone wants to admit.