A business rebuilt its website in 2022 with an agency. The retainer ended quietly in 2023. In the spring of 2026, somebody noticed the contact form had been sending enquiries to a mailbox belonging to a member of staff who left the previous October. The plugin powering the form had not been updated since 2022 either.
The site had looked fine the whole time. It was up, it loaded and the homepage photography was lovely. Nothing on the outside told anyone that the only route into the business from the internet had been quietly filling a dead letterbox for six months.
The website was never an IT decision. It was a marketing decision that turned into a server.
For many UK SMEs, that distinction explains the problem. The website is the most public system a small business operates: open to anyone on the internet at any hour, running code the business did not write, holding a database and sometimes collecting personal data. Yet it is often the internet-facing application with no named technical owner.
A business does not need a web team. It needs the site on the same asset register, patch cycle, access review and recovery plan as everything else.
What a website actually is
A corporate site is more than pages and photographs. Behind the visible design is hosting: a server, or a slice of one, running somewhere you have probably never visited. The hosting environment stores the site’s files and makes them available to visitors. Some businesses use cloud hosting as part of their wider Cloud Solutions approach; others use a straightforward hosting package that a competent provider manages well.
The site also has a content management system, or CMS, with a version number. A theme controls much of its appearance, while plugins add functions such as forms, bookings or search. Each is software maintained by someone, and each can have its own update, support and licence arrangements. The CMS usually works with a database containing the pages and other site information, plus a file store for images, documents and uploads.
Then there is the TLS certificate that enables the padlock in a browser, analytics and tag scripts, and any links to a payment provider, booking service or newsletter platform. The form handler sits somewhere in that chain too. Depending on how the site is built, the submitted information may be stored in the database, forwarded to a service provider or sent to a member of staff.
A modest marketing site can be genuinely simple, and a competent host may manage much of its infrastructure. But simple is not the same as self-maintaining. The site still has software that needs to remain supported, accounts that need reviewing and a backup that needs to be restorable. As soon as a visitor submits an enquiry, a CV or a booking, the site is also part of the organisation’s personal-data environment.
Performance belongs in the same conversation. Large image files, external scripts and the hosting tier all affect what visitors experience. Network Solutions are therefore part of the practical site picture, not just a question of design. Site performance can be measured rather than guessed, using indicators described in Web Vitals.

The logins nobody can list
The site may depend on a registrar account, hosting control panel, CMS administrator account, content delivery or firewall service, analytics property and search console verification. Add the payment or booking provider, email or newsletter platform, and the accounts used to buy plugin licences. The domain account belongs on the record too, but the question here is what operates the site once somebody arrives.
These accounts accumulate in the ordinary course of business. The original designer creates some. A second agency adds others. A marketing freelancer signs up for analytics. An office manager “just uploads the news”. A director may have registered an account years ago. One or more plugin licences may be renewing against somebody’s personal card.
The recurring discovery is that nobody can produce a complete list, and that some accounts are held by people who no longer work for the organisation. A retainer is a service arrangement, not ownership of the asset. When the arrangement ends, the business keeps the ongoing costs of anything it cannot reach.
It also matters which access level a person has. An editor who can publish a page does not need the same authority as an administrator who can change settings, install code or manage other users. A shared administrator login defeats every other control on the list: activity cannot be attributed properly, and access cannot be removed cleanly when someone leaves.
Named accounts and least privilege are familiar principles across IT. Microsoft’s role-based access control guidance offers one neutral illustration of how roles can limit what an administrator may do. The principle applies more widely than any one product: give each person an identifiable account and only the permissions their work requires.
The patch problem that nobody schedules
A CMS, its theme and its plugins are third-party code. Some are maintained carefully. Some depend on one person maintaining them in their spare time. Some are simply abandoned. An unsupported plugin sitting on a public website is an application without a maintenance route, and outdated plugins and themes have been a recurring path into small-business sites for years.
Before any of that matters, somebody has to know what is installed. A plugin inventory is unglamorous, and it is the difference between a maintenance routine and a hopeful one. It lists the version of the site software, the theme, and every plugin with its purpose and its current release. Two notes matter more than the rest: which plugins are no longer maintained, and which ones are no longer needed. Most businesses find at least one of each. A plugin that does nothing useful is still a plugin that has to be kept current, so the cheapest maintenance decision is often removal.
Updates can be awkward. Applying them automatically may reduce the time a known vulnerability remains unaddressed, but an update can also break a site, perhaps at the end of the working week. The answer is not a choice between “always update immediately” and “never touch it”. It is a controlled process: use a staging copy where practical, test the change, and have a rollback route if something goes wrong.
Cyber Essentials sets an expectation for high and critical security updates to be applied within fourteen days. That is a scheme requirement, not a law for a private business. Its wider relevance is the operational discipline it encourages: know what software is in scope, who maintains it and how updates are handled. The NCSC’s 10 Steps to Cyber Security provides a broader frame, while its Cyber Essentials overview describes the scheme.
The NCSC’s guidance on managing deployed devices is about devices, but reinforces a useful principle: deployment is not the end of management. An internet-facing service also needs an owner and a maintenance route, not a set-and-forget installation.
What flows through the site
Contact forms, quote requests and callback requests all collect information. So do newsletter signups, job applications with uploaded CVs, appointment bookings and complaints. A subject access request may even arrive through a web form that nobody thought to monitor.
For each route, trace what happens next. Does the submission stay in the site database, go to a form provider, enter a CRM or reach someone by email? Who reads it? Where is it retained? What happens if the process stops working? The question businesses often cannot answer is simple: when a form is submitted, what happens, in what order, and who is accountable if it stops?
The ICO’s right to be informed guidance explains why people need privacy information about what is collected, why it is collected and who it is shared with. An enquiry submitted through a website is still personal data the business may need to find. The ICO’s right of access guidance is relevant when considering whether the organisation can locate information held across its website and connected services.
Analytics and advertising tags are another part of the data flow. They may run before a visitor has made any choice, so check what the site actually loads and when. The ICO’s cookies and similar technologies guidance covers the relevant consent and information principles.
A private business is not bound by the public-sector website accessibility regulations described in government guidance. That is not a licence to publish an unusable site. A form that cannot be completed by keyboard or screen reader is a form that can lose customers.
A site that sends an enquiry directly into a customer relationship platform, or into a Business Central quotation process, is a different proposition from one that emails a mailbox. That integration needs the same governance as the website around it. A Dynamics 365 Jump Start can be part of a strategic discussion about where information should go and how the process should work.
What actually goes wrong
The site quietly goes down over a bank holiday weekend. Nothing is watching it, nobody has arranged uptime monitoring, and the first sign is a customer mentioning it on Tuesday.
A certificate expires at the weekend and visitors see a browser warning. To a customer, that warning can read as “this company is not safe”, regardless of how the certificate came to lapse. Elsewhere, a vulnerable plugin is used to upload a file and the hosting provider suspends the account. If the business email shares that hosting account, email may go with the site.
Sometimes the site still loads but the important part has stopped working. A plugin update breaks the enquiry form; nobody checks whether submissions still arrive, and the failure goes unnoticed for five weeks. An agency relationship ends and the hosting account goes with it, along with the only copy of the database. A staging copy remains reachable online with two years of old customer records inside it. The newsletter list turns out to be in an agency’s platform account, not the business’s. A former marketing assistant still has administrator access because nobody removes access they did not grant.

Each of these is survivable on its own. That is exactly why they go on for months, and why the bill arrives in one go.
If there is evidence of compromise, the organisation needs a response route as well as a technical fix. The NCSC’s incident management guidance covers preparation and response. If personal data may have been exposed, the ICO provides operational information about a personal data breach. The practical priority is to establish what happened, contain it and involve the people responsible for the business’s response.
Backup and recovery for a website
A useful website backup has to cover more than the homepage. It should include the database, uploaded media, theme and plugin configuration, and the redirect map that preserves years of search visibility. It should also include the DNS records needed to restore how the site is reached, without turning recovery into a separate domain-management project.
A host snapshot is useful, but it is not a recovery plan if it only reaches back seven days and nobody has ever restored from it. The only proof of a backup is a restore, done into a staging environment, with a person who has done it before. Hosting and cloud hosting choices should support that practical recovery route; Cloud Solutions should be assessed against what the business needs to restore, not just where files are stored.
The required recovery time depends on the role of the site. A brochure site may be able to wait a day. A site that takes bookings, applications or payments may not. Put the website into the continuity plan alongside other systems. Decide who declares an outage, who coordinates recovery, what the business tells customers while the site is down, and how the business will work in the meantime. Disaster Recovery is part of that plan, not a document to discover after a failure.
The cost side
Hosting may be on a tier the site has outgrown, or on a package that is more than the site needs. There may be a monthly agency retainer for updates and support that nobody can evidence, a separate CDN subscription, an SSL add-on, a backup add-on, plugin licences renewing against a personal card, overlapping analytics tools and a paid staging environment that is never used.
Technology Expense Management means reconciling what the organisation pays for against what it actually has. An inventory review is not a promise of savings. Its value is knowing what each line buys, who owns it and whether it still supports the plan.
The website also publishes the main switchboard number, branch numbers and opening hours. When a number changes or a branch closes, the site is often the last place anyone updates. That makes Voice and Connectivity relevant to website ownership too: a customer ringing a dead number on a current-looking page is a connectivity problem wearing a marketing costume.
A distributor or trade-counter site showing stock availability may read from the same IBM i system that runs the order book. In that case, the website integration has an account on the system, with permissions somebody should have deliberately granted.
A first month of ownership
A practical review can begin without building a web department. Start by finding the accounts and naming an owner for the service. Then make access, monitoring, recovery and spend part of the organisation’s normal IT routine.
- Pull every account into one list. Record what it does, who holds it, how it is accessed and who is responsible for it.
- Bring hosting and CMS administration into organisation-owned accounts. Enable multi-factor authentication on both.
- Replace shared administrator logins with named accounts. Give each person the lowest role that lets them do their job.
- Switch on uptime monitoring and certificate-expiry alerts. Make sure alerts reach somebody who will act on them.
- Take a full backup and restore it into staging. Record what was restored, who completed it and whether the site worked afterwards.
- Set a monthly patch window. Include a staging step and a rollback route, and record anything that cannot be updated.
- Check every form’s destination. Confirm that the privacy notice and cookie behaviour match what actually happens on the site.
- Add the site to the asset register. Name an owner, set a review date and read the relevant invoices alongside the inventory.
Managed IT Services can be a bolt-on execution service for monitoring, the patch window and periodic review. Helpdesk Support can provide the documented route when the site breaks, a certificate is expiring or an access change is needed at 8am. A support desk cannot fix a website it has never been told about. That is the same point as the inventory.
A named owner should also be able to say whether the service is still right for the business. That may mean reviewing hosting within the broader Network Solutions environment, confirming who performs the work and setting an agreed route for escalation. The purpose is not to make every website decision an IT project. It is to make sure decisions have an accountable person.
A website also ages in a way that is easy to miss, because nothing on the outside changes. A design that looked current in 2021 still looks like a website. Underneath, the software running it may be two major versions behind, the plugins may be pinned to that older version to keep it working, and the person who built it may have moved on. When the rebuild finally happens it tends to happen in a hurry, after a failure rather than before one. The alternative is a line in the register: what the site runs, who supports it, when that support ends, and the date by which a decision is needed. That is the same discipline as any other piece of equipment the business depends on, and it costs nothing until the decision actually arrives.

Who owns it now?
The website is the part of the business anyone can reach at any hour. Whether anyone is accountable for it is a fair test of whether anyone is accountable for the rest. The question is not “who built the site?” but “who owns it now?” The answer has to cover hosting, code, access, data, recovery, cost and replacement.
A Fractal IT Director provides the strategy layer: the decisions about risk, priorities, investment and how the site fits the wider technology estate. A trusted network of specialist channel partners can add capacity and specific technical execution. Bolt-on services, including Managed IT Services, Helpdesk Support, and Voice and Connectivity, support that strategy. They are execution services, not substitutes for the decision.
For the wider group of services, the gateway is evestaff.co.uk.
A website is not a brochure. It is a server with a login, and the business should know who holds both.
SEO keywords:
Cloud Solutions, Disaster Recovery, Dynamics 365 Jump Start, Helpdesk Support, IBM i Management, Managed IT Services, Network Solutions, Technology Expense Management, Voice and Connectivity, IT Consultancy, Fractal IT Director
website management, website security for SMEs, content management system patching, website backup and recovery, website ownership
Join The Discussion