It did not happen suddenly. Two years ago a page took a moment. Now your staff click it and go make coffee. Somewhere along the way everyone stopped mentioning it and started working around it.
Slow is the symptom people notice first and take least seriously, because there is no outage to point at. But a system that gets steadily slower is telling you something specific about what is happening underneath it.
The four things it usually is
Almost every case comes down to one of four causes, and they have different fixes and very different costs.
Your data got big and the system was never told what to expect. This is the most common cause by a distance. When a system is new it holds a thousand records and finds any of them instantly. At two million records, the same lookup means reading the whole pile every single time. The fix is an index — essentially a lookup table that tells the system where things are. Adding one is often minutes of work and can turn a nine-second page into a fast one. This is the good outcome, and it is worth ruling in or out first.
The system asks the same question hundreds of times per page. A screen that shows fifty orders asks for the customer name of each one separately, one after another. Fifty small delays become one long one. This is a well-known pattern with a well-known fix, and it is usually a day or two of work per screen.
You have outgrown the server. Old shared hosting gives you a slice of a machine you share with dozens of other websites. When one of them gets busy, you get slow, and there is nothing you can do about it. If your slowness comes and goes without any pattern you can find on your end, this is a likely cause.
The software underneath is simply old. The language your system is written in has had years of performance work done to it. Moving from a version released in 2019 to a current one routinely makes everything meaningfully faster, with no changes to your code at all. This is the least glamorous fix and often the largest single improvement.
Which one is yours
You can get part of the answer without any technical help. Ask your host which version of the underlying software the site runs on and whether it still receives security fixes. If it stopped being supported years ago, you have found at least one of your causes, and probably the cheapest one to fix.
The rest requires someone to measure. Not to guess — to actually put timing on the pages your business uses, in order of how much each one matters. The output is a ranked list: this page takes eleven seconds, here is why, here is what it costs to fix, here is what your staff lose to it every week.
The reason to measure rather than guess is that intuition is bad at this. The page everyone complains about is often not the one costing the most, and the fix that sounds expensive is regularly the cheap one.
What slow is actually costing you
It is worth putting a number on this, because slowness gets tolerated in a way that outages do not.
If a screen your staff use forty times a day takes eight seconds instead of one, that is roughly five minutes a person per day. Across six people that is half an hour daily, or about three working weeks a year. That is a real cost and it is invisible because it never appears as a line item.
On the customer-facing side the arithmetic is harsher. People leave slow pages. They do not tell you they left. Whatever you spend getting people to your site, a portion of that spend is being thrown away at the door.
And there is a third cost that is easy to miss: slow systems train your staff to avoid them. They start keeping spreadsheets on the side because the real system is too painful. Now your data lives in two places and neither is right.
Why it accelerates
Slowness compounds for the same reason it starts. Your data keeps growing, so the lookups that were merely inefficient become genuinely expensive. Nobody has removed anything, so pages accumulate more work over time. And because the system has no safety net for testing changes, performance work gets postponed in favor of anything more urgent — which is everything.
The version of this that ends badly is the one where slow becomes down. A system operating near its limit has no headroom. A modest spike in traffic, a slightly larger report, one extra person logged in, and it stops entirely. The first real outage is usually a surprise to everyone except the person who has been watching the numbers, and in this situation nobody is watching the numbers.
What fixing it involves
The order matters, because the cheap wins come first.
The audit measures and ranks. Five business days, $2,400 fixed, and you get a written report you own — plain language, no jargon, with the cost of each fix next to the size of each problem. Quite often the first two items on that list are an index and a version upgrade, which together cost a small fraction of the audit and remove most of the pain.
If the cause is the server, moving is a separate piece of work. Modern hosting is faster and frequently cheaper than the old shared plan you are on. Your web addresses stay exactly the same, so nothing you have built up in search is affected, and the switch happens at night with a tested way to undo it.
The full price list is here, including what is free. There is no version of this where you have to book a call to find out what something costs.