Technology has made it too easy to create content. As a result, we have too much marketing content for engineers.
From search engines that make research easy, to ever-present smartphones that take photos, to instant sharing on social media, technology has contributed to an avalanche of content. The engineers we seek to reach are trying their best to filter out the irrelevant messages so that they can get on with their work.
Marketers now complain that it is harder than ever to get their messages in front of technology buyers. They are watching their email open rates decline while they pay ever more to distribute their content on social media.
The Marketing world's recent obsession with content has dug us into this hole, but technology may also provide the way to dig ourselves out. We just need a little help…
I was at Content Experience 2017 in Toronto this week and heard a lot about AI and how it will work with marketing automation. We have a little experience with it ourselves. As you’ll see in this post, we’ve learned a lot about how to leverage machine learning to improve the experience of content discovery for our target audience.
This post will cover my take on:
- Why content clutter is breaking the internet for engineers
- How machine learning and artificial intelligence works to make better content recommendations
- What this will mean for engineering marketers
Why is content production breaking the Internet for engineers?
Engineering marketers aren’t competing only with other marketers for engineer’s time. They are competing with every other stream of content that might catch an engineer’s interest.
According to Eric Schmidt, in 2010 people created as much information in just two days as all of humankind produced up to 2003.
At Conex2017 Yoav Schwartz added that next year, it will only take humanity 10 minutes to produce that much data. That’s a lot of competition for engineer’s time.
And it gets worse.
Here’s a staggering fact. In 2014, there were 1.8B photographs being uploaded to the Internet every day. That just wasn’t possible in the olden days before smartphones.
There used to be technology constraints. With film cameras, you could only take 24 pictures. Now with smartphones, there is virtually no limit.
You might be thinking, “Surely marketing content for engineers isn’t exploding like that.”
This post, Content Marketing to Engineers is Tough. And getting Tougher makes the case that engineering marketers, like everyone else, are investing more in content.
Some of that content isn’t even seeing an audience. According to Sirius decisions, 60-70% of B2B marketing content goes unused. To address this issue, many technology marketers are investing more in content distribution, which is a good idea. But only until everyone else does that too.
As Jay Baer pointed out, (again at Conex 2017 which was an excellent conference btw), you can’t “out volume” mediocrity.
Technology Just Might Hold the Answer
Like all content, there has been an explosion of video content and self-published books over the past few years. But somehow, Amazon and Netflix have made it manageable. They show you recommendations for content that are super relevant.
(What I see when I log into Netflix. I have no idea how Bones got in there. I may have watched an episode of Angel once..)
The algorithms Netflix and Amazon use to make recommendations are highly guarded secrets, but we know in general how they function. Their excellent recommendations are made by correlating what other people watched and engaged with (as measured by ratings and by how long they watched) and using that information to predict what you will like.
For example, let’s say I liked House of Cards (true). Let’s say that other people who liked House of Cards also liked Breaking Bad. If I hadn’t seen Breaking Bad (not true), then Netflix would recommend Breaking Bad to me.
This kind of recommendation is more complex than a simple contextual based match like you might see in a Google Ad. In many ways, it is more powerful.
At ENGINEERING.com we have been using leading-edge technology to power our content recommendations for years. In late 2015 Amazon web services released a beta version of their Elastic Search features which provided contextual matching. We implemented that service in our own recommendations. It now drives more than 4,000 clicks every week, building deeper engagement with our audience.
This year Amazon released a machine learning platform that allowed us to build recommendations based on what other engineers liked. It looks like it is doing very well in the early going.
ENGINEERING.com machine learning recommendations at the end of each article
AI - The Future of Marketing Automation
Artificial intelligence and machine learning are now becoming more accessible to software developers in every industry. If ENGINEERING.com can build a machine learning recommendation engine, then you have to expect that this sort of functionality will soon be in your marketing automation tools.
I predict that you will soon be able to use these tools to deliver customized content recommendations at scale. This will allow you to deliver unique nurture flows to each visitor to your web site, leading to higher conversions.
In the meantime, given that marketers are dedicating more budget to content than ever, your best path forward is to direct more budget towards creating only the very best content for your marketing. You can’t “out volume” mediocrity.
Until next time,
John
