You know that feeling of relief you get when you finish a big marketing project? You did everything right, the bosses seem satisfied, and you feel like you can kick back and relax. I remember that feeling, as it lulled me into a false sense of security. Everything seemed great, I had sorted out all of our problems, and everyone treated me like I was the company rock star.<\/p>\n
Until six months later\u2026<\/p>\n
The honeymoon phase had ended. Suddenly, management no longer felt satisfied. All those feelings of relief suddenly turn into sensations of dread.<\/p>\n
Has this ever happened to you? It all starts with the purpose of the project in the first place. Did you ask the team about the project’s purpose? I mean, really ask what they really want? Let’s say, for example, that the site’s purpose had to do with selling more products. You might know that management wants to sell more products, but do you know what “more” actually means to them? You see, unless you can point to a quantifiable number that defines success, you have doomed your project to failure from the start.<\/p>\n
Alas, if you have experienced this, then you know that we did not do that research before we launched our sites. Six months passed and we found ourselves in this uncomfortable situation. What do we do? You could scrap everything and start again, but that amounts to an enormous loss of time, effort, and money. Not to mention the fact that it virtually admits to management that you didn\u2019t completely understand their goals when you started. Honestly, though, unless you completely missed the mark (unlikely if you\u2019ve made it a few months without too much commotion), you probably have a pretty decent work product that just needs a little tweaking.<\/p>\n
In this situation, you probably just need to make a few iterative changes to bring the campaign more closely into alignment with what management envisioned. A\/B testing, or “split testing” as it’s sometimes referred to in our industry, provides one of the best solutions to accomplish that goal. A\/B testing refers to trying different variations on a single controlled site (or piece of a site) and seeing which variant creates the best reaction.<\/p>\n
Unfortunately, A\/B testing offers its own unique challenges. For those that haven’t tried it before, an A\/B test [without the proper process] may go something like this:<\/p>\n
\nI got a lot more from this talk then I thought I would! Thx @mpmike<\/a> #wcbos<\/a> #ducks<\/a> pic.twitter.com\/E4vrv61Zeo<\/a><\/p>\n
— Meagan O'Brien (@meaganaobrien) July 23, 2017<\/a><\/p><\/blockquote>\n