Replace your website when keeping it costs more than replacing it. That usually shows up in one of four ways. The software can't be patched anymore. The site stops customers from buying. Every change has to go through a developer. Or it runs on a machine in your office that nobody should be relying on. A dated look, on its own, almost never justifies starting over.
Most of the advice you'll find on this lists the dated look first. I think that's backwards, so I'll explain with a project I finished this year.
The site that had to go
The client sells custom-made products online. When they came to me, they had two online stores built on CodeIgniter and PHP 5, plus an internal intranet written in classic ASP. All three ran on Windows machines inside their office.
PHP 5 stopped getting security fixes at the end of 2018. Classic ASP was replaced by ASP.NET in 2002. Neither one was going to be patched again, no matter what turned up.
What turned up was a defaced homepage. Someone broke into one of the stores and replaced its front page, and customers saw it. I wasn't the one who cleaned it up, and I don't know how they got in. What I do know is that the software it ran on could no longer be patched, and the owners knew it.
The break-in forced the decision. The stores had been costing them money for a while before that.
Their products have to be built to the manufacturer's requirements, and the old stores couldn't handle the options. A customer would choose a combination that couldn't be built, or couldn't find the option they needed at all. The order then turned into a refund, or into a member of staff on the phone walking the customer through it by hand to save the sale. By the owners' rough estimate, that cost them several thousand dollars a month.
On top of that, every product change went through their old developer. Adding a product or retiring one the manufacturer had dropped meant sending a request and waiting.
Signs it's actually time
Here are the four signs from that project, in the order I'd check them on any site.
The software can't be patched. Find out what your site runs on and whether it still gets security updates. If the language or the CMS version is past its end of life, every new vulnerability stays open forever. Upgrading in place is sometimes possible. Once a site is two or three major versions behind, an upgrade is usually a rebuild with extra steps.
It's turning customers away. Look for orders that fail, checkouts people abandon halfway, forms that error out, and options customers ask for that the site can't offer. Ask the people who answer your phone, too. If they spend part of every day finishing transactions the website should have finished, that's a monthly cost you can put a number on.
Every change needs a developer. If you can't change a product or a page yourself, you pay for that in invoices and in delay. A product that takes two weeks to go live is two weeks of sales you didn't make.
It lives in your office. A server in a back room depends on your office's power and internet connection, and on a single machine. It also ties your staff to the building. Moving to proper hosting is often possible without a rebuild, but on an old enough stack the move and the rebuild end up being the same job.
Signs that usually aren't enough
A site that looks dated or loads slowly is worth fixing. It usually isn't worth replacing for that reason alone.
Slow pages are most often oversized images and third-party scripts, and images are the usual culprit. Phone layouts can often be fixed in the existing theme. A homepage that doesn't convert is usually trying to do too many jobs, and a form that loses leads usually asks for too much. Each of those is usually a few days of work.
If those are your only problems and the software underneath is current, fix them and keep the site. If they sit on top of the four signs above, they're one more reason to start over.
What the rebuild changed
We replaced all three systems in two stages. The first store and the intranet went live after about twelve months. The second store followed about six months later, in May of this year. Staging it meant the business never had to switch everything over in one weekend, and the second store's build could learn from the first.
The product options were the center of the project. On the new stores, options are checked against the manufacturer's rules as the customer picks them. Combinations that can't be built are blocked before checkout. The price updates as the customer chooses, and the finished order goes to the manufacturer already in the format they need. The old stores couldn't do most of that.
The client's staff now add and update products themselves. Nobody waits on a developer to put something new on sale.
Everything moved out of the office to cloud hosting, intranet included, so staff can deal with customer problems from anywhere.
The refunds and the rescue phone calls were the cost the owners felt most, and that time now goes into selling to new customers. Since the first store relaunched, combined sales across both stores have averaged 34% above the same month a year earlier. I won't claim the rebuild did all of that. They added products in the same period and dropped some the manufacturer no longer makes, and for the first six months one of the two stores was still the old one. Customers being able to order what they actually wanted accounts for a lot of it.
What a new website costs
These are the ranges I work to.
| What you're replacing it with | Typical range |
|---|---|
| A brochure site, 5 to 10 pages | $3,000 to $8,000 |
| A custom-designed site your staff update through a CMS | $8,000 to $20,000 |
| A site with custom features, such as booking or a client portal | $15,000 to $40,000 |
The top band overlaps with custom software on purpose. Once a site takes orders with business rules behind them, or runs part of your operations, it is software, and custom software is priced differently.
For comparison, vendor pricing guides published this year put a template brochure site at $1,500 to $8,000 from a freelancer and $5,000 to $15,000 from an agency, and a custom site with a CMS at $5,000 to $15,000 freelance and $15,000 to $40,000 from an agency. They come from vendors, so treat them as a rough picture of what you'll be quoted.
Budget for running costs as well. Hosting and upkeep for a small site run roughly $50 to $200 a month by those same guides, and more for a store or anything your staff log in to.
Then compare the total with what the old site costs you. A store losing several thousand dollars a month to failed orders is paying for a replacement whether or not it buys one.
How to decide
If you're not sure, answer these four questions about your current site:
- Does the software it runs on still get security updates?
- Can customers buy or book everything you sell without calling you?
- Can your staff change products and pages without a developer?
- If the machine it runs on died tonight, how long would you be offline?
One "no" to the first question is enough to start planning a replacement. If the answers to the others are costing you money every month, you already have the number that tells you how much a new site is worth to you.
