Most smaller businesses don’t think much about where their website lives on the internet. Often that’s left for the website designer to sort out.
Sometimes people buy hosting because someone saw an ad with a big cheap price tag prominently displayed, and thought “that’s cheap, it will do”.
In general I find people spend way more time and effort choosing colours and fonts and picking out their domain name, than they do thinking about the hardware that makes their website work.
Full disclosure: we sell website hosting to our customers. Some people prefer to get the hosting bundled with other services, but that’s not the main reason we sell web hosting. The main reason we sell web hosting is that dealing with randomly chosen cheap web hosting is very often very timeconsuming and frustrating.
Here’s the story of a website – not developed by us – that we were asked to help with. The original problem was:
The website was producing random errors when the site owner edited a page.
The errors happened regularly, but were very hard to reproduce – sometimes editing a page would cause information to disappear from the website, sometimes it wouldn’t.
We spent hours testing different kinds of changes, to try to work out if changing THIS field would cause a problem. No? OK, make a note, try changing THAT field. Check the code, check the database. Make a note.
This kind of testing is expensive because it takes so long – and these errors were extremely frustrating for the website owner, who never knew when the site would have a hiccup and lose information that she had carefully entered into it.
The website that could not be copied. To try to determine what the problem was, we tried to install a copy of it at a different location, so we could safely test without creating problems with the live website. But the copy would not install properly. It kept producing worrying new errors, that we did not see on the original website! It was all very frustrating.
Both of these problems eventually turned out to be the SAME problem. And that problem was…. The hosting was really slow.
I’d noticed when making changes that the website was often slow to respond, and I suggested that a change to faster website hosting might please website visitors and offer a better experience.
But I had not realised just how much impact this was having on the whole system. When the website owner tried to update her website, if it was a busy time, there was a good chance that the website would lose information rather than saving it.
When I tried to copy the entire site to a new location for testing, the connection would drop before everything had been retrieved, so that the copy of the website I took was not complete. Because the copy was missing vital information, it could not run properly and showed a series of weird errors.
This led us to the alarming conclusion that this site was also … The website with no backups.
Although the website had a nice friendly-looking ‘backup’ button, and the site owner had been regularly hitting the button and storing the downloaded file in a safe place, the backup was worthless. Because the website was so slow, the backup routine was simply not catching everything.
So, the website, which had had months of time and effort and not a little money invested into it, could be lost very easily. If the cheap hosting company had gone bust, if the website had been hacked, if the hardware had failed – everything would have disappeared.
This isn’t the only case where I’ve seen this. The slow buggy website as a result of cheap bad hosting is alarmingly common. It can cause security problems too – I had one this year that was on very cheap hosting, where I found a pile of secret forms that looked as though they were trying to collect credit card data. I think they were installed by a hacker who had probably been using another account on the same shared server.
And all of this because nobody thinks about hosting. Nobody considers that a website which started out with a handful of visitors and maybe thirty pages of information, now had hundreds of visitors looking through and interacting with thousands of pages, and that the needs of the site had changed.
Is my web hosting up to the job? If your website is a local business site, aimed at small numbers of local customers, and probably you will never have more than 4-5 people looking at it at once, then your webhosting needs are pretty minimal.
If your website doesn’t change much, if there are no forms to fill in, no database or dynamic personalised information, then the demands you place on the webserver are small. Even if the server is a bit busy, probably your website will be perfectly useable.
But if your website has a content management system or a shop – if your website might get popular – then your web hosting should be planned so that you can easily scale it up.
Even web developers tend to assume that hosting should be cheap. But bear in mind that hosting doesn’t just need to go ticking along at the good times. It also needs to have a plan to cope with increased popularity, a team of engineers who can stand things back up when it falls over – even if it falls over at 3am on a Bank Holiday Monday. It needs backups, and the backups need testing. None of these things is cheap, and you don’t set them up once and then never look at them again.
Next time you need hosting for a web project, ask the hard questions : how is this backed up? How do I test that? What happens if there’s a hardware failure? How many people can I reasonably expect to see my website at once? Are the backups kept separately from the website? If the entire datacenter has a problem, are my backups inside it? Ask to see another site using similar hosting, and find out how fast it is – the Pingdom Speed Test is handy.
You’ll almost certainly end up paying more for web hosting than you think you ‘should’. But if you do it properly the first time, you’ll never know how much time and money you just saved yourself. And that’s a good place to be.
