Do You Know Why Your Marketing Content Is Working?

• July 28, 2017

Most large companies have embraced data-driven marketing by now. (If your company hasn’t, I suggest you give it a big hug right now.) But along the way to Big Data Nirvana, a few problems arose. A big one is understanding the difference between correlation and causation. For example, on an e-Commerce site, do you know which web page is most associated with purchase? It’s the Thank You page. You probably guessed that great Thank You pages don’t cause more purchases.

So, yeah, that’s a silly example, but we make similar mistakes all the time. Every time I work with a client to help them figure out their best content, they fall back on the same numbers–which content led to conversion. So, they know to throw out the Thank You page and the shopping cart page, but when they look at the list, guess which pages show up? The best-selling products. OK, but is that because the products are better or the content is better? We don’t know.

Even if we could tease that apart, it’s still unsatisfying, because the reason you want to identify the best content is to make more of it. But how helpful do you think it is to point at a successful page and tell your content writers, “Make more like this one.” One writer slowly raises her hand and quickly asks, “Like that page in what way, exactly?'” At that point you give them all a blank look and start to drool just a little.

You don’t know. You have no idea. You might know that it worked, but you don’t know why. How can you answer this reasonable question? Enter artificial intelligence.

You need one AI technique to tear apart the page–text analytics. You need data analytics to identify which pages have the best outcomes (conversion rates, inbound links, social shares–whatever you think identifies success). You need machine learning to suss out which characteristics seem to be shared by pages that are successful.

Now you can answer the writer’s question because you know exactly in what way the should design the new pages. You know whether pages with bullets work better than ones with streams of text. You know how many images are too many. You know if using brand names at the top of the funnel is a turn off.

You know a lot. And the more pages you look at, and the more characteristics of those pages you look at, and the more activity on your website, the more you learn. Artificial intelligence isn’t the future. It’s now. It isn’t magical–it’s very practical. If you are not doing it, maybe the competitor is answering questions that cause you to drool.