Your Website Will Probably Fail

It took me a while to get the courage to write this post — that is, to tell you that your website, new or redesign, will probably fail. In fact, there’s a 65-80% chance. But wait, there’s more! It may fail multiple times! That’s right, there is a reasonable chance you will experience multiple failures after your website goes live. Now, before you get up in arms, let me explain.


Let’s start with describing what a failure is, with context. In our industry, if a website doesn’t make your business better — doesn’t make your business grow — much less makes it worse, then it’s a failure. In fact, it’s so likely to fail one or more times out of the gate that we plan for it. If you’ve worked with a web designer or digital marketing agency in the past, that may not be your experience. We hear a lot of stories about one-and-done projects. That is, you pay someone to create a website for your company, they work with you to create a pretty design, they put some SEO in place to help you rank in search results, and maybe even offer a few more value-added services. But, they forget something critical in this process — testing.


Let’s clarify what we mean here. Testing in our business can mean a couple of different things. Maybe what comes to mind is the testing that happens before the site goes live. Yes, it’s really important to test a number of technical things before you unleash the site for public consumption. That may include testing dependencies, security, linkage, compliance, or any number of elements related specifically to your deployment. This is a requirement, and any good provider will make sure these steps are part of their pre-deployment process.

What about after the site goes live? Maybe some more testing of the technology is required to make sure the site is performing as it should in the production environment. There may even be a few things that need immediate testing in production that couldn’t be tested in a staging environment — all legitimate steps in most design-develop-deploy methodologies.

The testing we are talking about here is something quite different. And, it happens after your site is live.


The reason we plan for your site to fail is because there is such a high likelihood that we will, at least initially (Disclaimer: Not many professionals will tell you this). Why wouldn’t we be prepared for failure, knowing it’s likely, rather than react to it after it happens? That’s good ole’ common sense, right? But, before you get terribly concerned that your website is going to fail, let’s go a little deeper and determine what that really means.

These days, almost all consumer products are designed in such a way that there is preparation for failure. When you buy a new consumer product, chances are there were several versions that failed in test groups before you got your hands on it. With the introduction of iterative software development methodologies, we build things in small increments, test them, and iterate the things that need to change. In other words, we invest a little (or just enough) on multiple smaller things, figure out which works best, and then double down after we get the answer. It saves us from doing a tremendous amount of work before getting feedback. In turn, it saves you from having a product that you’ve gone all-in on where there is a likelihood it may not work as it was intended. And, it takes care of the idiom that the only constant in life is change. So, when we design and develop a website for a customer, we expect there to be change. Why? Because no one knows how consumers are going to react to it. There are plenty of freelancers and agencies that tell you, because of their vast experience, they know how to make your site successful the first time. They will build it, and people will come! Sorry, but they won’t unless you get very, very lucky. I’ll say it again, no one knows how consumers will receive your website until there is proof in data.


Testing can be complicated, but the explanation doesn’t have to be. Building a business is hard — you already know that. Testing takes planning and preparation, a little extra work upfront, and the willingness to listen and make changes when you get the results. The entire idea is to move the needle, and sometimes that takes longer than you expect, but patience, in this case, is a virtue. The alternative is to construct your new site or redesign in a way that doesn’t allow for testing, or more accurately, doesn’t make provisions for comparing tests so that you know what’s working best. Why would anyone do that? It is like flipping a coin one time when getting heads is imperative. Wouldn’t you feel much more confident knowing you could flip a coin as many times as you needed to in order to get heads? Of course, you would. That’s the luxury that testing provides.

In consumer website testing, we refer to this type of testing as A-B, A/B, A/B/n, split-testing, or multivariate testing. What that means is that you allow the consumers of whatever it is they are consuming to have multiple experiences and test the results. The potential population of consumers (in this case, visitors to your site) is unknowingly split into multiple groups — for the sake of simplicity, let’s say Group A and Group B (it’s worthy of note that this type of testing is not limited to just two groups). After a specified length of time, we analyze the data showing how each group interacted with your website, where conversion happened the most, and why. With any luck, we achieve good results from one or more of the tests and can take what we learned and apply it. This is a much over-simplified explanation of multivariate testing but illustrates the basic principle.

A/B/n testing is not exclusive to just websites, and it has become commonplace in many consumer product industries. Most professional web design agencies should employ testing to better ensure positive results for their customers.

ab testing


Remember that we said 65-80% of sites fail. In this case, we’ve done some work to make sure our testing is diverse enough that we have a much better chance than just one coin flip, or the one-and-done approach. But, even that does not guarantee success in the first go of things. Also, remember that we talked earlier about iteration. If, after running our planned tests, we haven’t rendered the results we believe we should be getting, we take the best of what we learned, tweak it, and put it back in the hopper for more testing.

Let’s be more specific. If we built 4 versions of a website landing page that we expect to garner certain results, and none of the versions are working quite as well as we thought, we analyze the data. We determine what’s working best by comparing several data points including bounce rate, user behavior, heat maps, click-throughs, and other user interactions. From there, we can decide if there is value in a single design or if we need to take the best of multiple designs and aggregate them into a new design to test further. Either way, we are iterating on the feedback that we’ve gained from the testing.

This might seem time consuming or complex, but it only adds a small fraction of the time to do the work up front rather than failing without preparation. It should be part of every agency’s process. And more importantly, it is THE ONLY WAY to look at a customer and tell them that we can assure results. The one-and-done approach will never stand up to that guarantee, and only by being very lucky will your website result in optimal return on investment on the first try.


This is a difficult question and there is no good answer. Without disparaging any other designer or agency, many come from pure artistic design backgrounds, or they haven’t experienced the rigor of iterative product development. The concepts here are somewhat fundamental, but without a comprehensive understanding, it may be daunting to some.

When discussing your website project with a potential provider, ask them how they test. Now that you know the difference, make sure that they have a process for A-B or multivariate testing, and that they include the ability to iterate your design based on test results in the initial production deployment or go-live. The time invested up front to cover this pays dividends on the back-end, and is the only way to ensure that your site is successful — meaning, that it makes your business better and not worse.

Working with a professional is key in making sure that your web technology investment is rewarded. A good digital strategy is more than a pretty facade. It requires patient planning, thorough analysis, market knowledge & awareness, and rigor in deployment and testing. We would love to talk with you about your website project. Please call us at 912.236.6003 or reach out at hello@delegal-digital.local

Like this POST?

Share on facebook
Share on Facebook
Share on twitter
Share on Twitter
Share on linkedin
Share on LinkedIn
Share on pinterest
Share on Pinterest

Leave a comment