Websites are usually not broken; they are simply frozen in that year.
A website built in 2019 still opens and works normally today. The pages aren't messed up, it isn't slow, the server is still being renewed, and the boss thinks it's fine—spending a little money each year to keep it running is no problem.
But lay out everything inside and take a look: the product page still has those 12 models, 4 of which were discontinued two years ago; the newest case is that 2021 project; the team introduction still displays photos of three people, two of whom have already left; the contact information is still the landline from back then.
The website isn't broken. It just stayed in 2019 and hasn't moved since.
What customers look at first when they walk in isn't the introduction—it's "is this company still alive?"
A procurement person looking for suppliers opens your website, and the first judgment they make is not "Do they write well?" but "Is this company still operating normally?"
They will look at a few places: when the last news item or update was posted, what year the case studies are from, whether the product page has any new products from the past two years, and whether the contact number is a mobile phone or a landline. If even one of these is obviously three or four years old, in their mind they have already crossed you off. They will not tell you why; they will just stop replying to your messages.
So the website's first function for new customers is not to introduce, but to verify whether this company is still operating normally. If this is not done well, no matter how beautifully written the rest is, no one will read further.
Technology depreciates slowly, while content depreciates quickly.
The pace of these two things is completely different.
Technical depreciation is slow: a website built five years ago can still open today, the server is still running, and the domain name is still resolving. So it is hard for bosses to feel that “it is aging,” because it really has not broken.
Content depreciation is fast: the industry changes once a year, products iterate every six months, several customer case studies accumulate in a year, and staff turnover can happen at any time. These things expire far faster than technology ages.
And that is exactly where the problem lies—technical problems trigger alarms: pages cannot open, errors appear, or the site is hacked, and someone will contact you immediately; content expiration never triggers an alarm. It just quietly makes everyone who comes in lose interest.
So the real lifespan of a website is not its technical lifespan, but its content lifespan.
After AI started giving answers, the cost of content becoming outdated was amplified.
Customers will still open a website and judge for themselves; AI won't.
When AI is asked "which suppliers in this industry are reliable," it looks for material it can verify: product specifications, delivery capability, publicly available case studies, and qualification information. To it, a website that hasn't been updated in three years looks much the same as a store that has shut its doors—not because you're not good, but because it can't read what you look like now.
In the past, outdated content cost you a little in customer experience; someone might still be willing to call and ask. Now, outdated content means you simply don't make the shortlist—and you will never know, because the people who overlook you won't come and tell you.
Maintaining a website is not about reposting two industry news articles every month.
What many companies understand by “regular maintenance” is simply putting two industry news items into the news section every month. Reposted content does not help resolve any specific customer judgment, nor does it constitute material that AI is willing to cite; it only makes the page look like it is moving.
In fact, only five categories truly need to be kept up to date:
Product models and specification parameters, price ranges or quotation methods, customer cases that can be made public, delivery and service capabilities, and contact information and team information.
Update when these five categories change; when they have not changed, there is no need to force something together. Content is supported by facts, and forced updates cannot support any judgment.
The more cost-effective approach is to manage it as an asset, not deliver it as a project.
Building a website is a one-time project; managing content is an ongoing practice. The difference between the two is this: a project is delivered when it is done, and once delivered, no one is responsible for it; an asset needs someone accountable, a place to accumulate, and a record of changes.
For a company that has been in business for ten years, not much can truly be retained. Products get replaced, people leave, channels change—but the product materials, project records, customer cases, and technical documentation accumulated over these years are exactly the parts others cannot copy.
The website is merely the outlet for these things. No matter how well the outlet is built, if there is nothing inside, it is empty.
A website may technically survive for ten years, while its content may only last two. How much longer it can still be used depends on when you last seriously updated it.