You’ve spent years developing your personas and you’re tasked with building a content strategy that will meet the distinct needs of each one. This should be easy. In true Pygmalion fashion, you’ve shaped their personalities, and you’re even more familiar with the intimate details of these prototypical models than you are with your own family members. You’ve constructed their names, jobs, loved ones, joys, fears, habits, aspirations, and social activities. But here’s the rub: while personas can help shape your initial strategy, they aren’t enough to keep your audience engaged in the long term.
Image attribution: Musée National du Château et des Trianons
Why? Abstract archetypes or personas don’t buy our products and services or read our publications. People do. Even if you’ve done a great job with persona creation, you can’t provide the right information to the right person at the right time relying solely on archetypes of your readers or buyers.
But, today’s digital natives demand this. According to Janrain and Harris Interactive, 74 percent of consumers get frustrated with websites when content appears that has nothing to do with their interests. But, 60 percent of marketers struggle to personalize content in real time, according to a study conducted by Adobe and the Direct Marketing Association. In most cases, engagement metrics such as time on site, number of pages viewed, bounce rates, returning visitors, and email CTRs reflect this struggle.
Paul Roetzer, CEO and founder of PR 20/20 and the Marketing Artificial Intelligence Institute, refers to this as the “Amazonification of marketing.” “People love the convenience of a personalized experience. They don’t want to be presented eBooks in which they have little interest. They want content that appears to have been written specifically for them that addresses their pain points. And, they have come to expect this without even realizing it.”
But, when it comes to content marketing, most marketers don’t have the ability to respond in real time and at scale. This requires artificial intelligence.
It’s Time to Make Content Personal
To address this challenge, Skyword built Skyword Personalized Recommendations (SPR). SPR (pronounced spur) is an artificial-intelligence-based engine that delivers personalized on-site and email content recommendations to increase engagement and conversions. In November, the Content Standard deployed SPR on site—and, shortly after, in emails and our subscriber newsletters. You can see it in action with the “Recommended for You” widget that appears within this article.
I was excited to apply personalization to help with our own engagement challenges. Over the past several years, we’ve created thousands of stories designed to inspire marketing transformation and help brands navigate the ever-changing digital landscape. Our stories perform well in search, and our pageviews grew 180 percent in the past two years alone. We add more than 1,000 new subscribers each quarter. However, our engagement is relatively stagnant. Despite an abundance of great content, we are challenged with making it easy for people to find relevant stories that are buried deep within our site.
The beauty of SPR is that it is built with an ensemble algorithm to deliver recommendations and “learn” from machine learning algorithms, such as collaborative filtering. Collaborative filtering is a way of making automated predictions about a user by collecting preferences or taste recommendations from other users. This type of approach is considered the most accurate and effective method of personalization.
We deployed content recommendations on the Content Standard in sidebar and below-article locations. We also started including personalized recommendations in our weekly subscriber newsletter.
Here are some results thus far:
- On site: Averaging 6X lift in CTR for machine learning recommendations vs. random
- Emails: 369 percent increase in CTR for emails with personalized recommendations in A/B testing
- Subscribers from emails: 5X increase in subscribers for emails with personalized recommendations in A/B testing
- Newsletters: Overall 116 percent increase in CTR to site for newsletters with personalized recommendations vs. those without personalized recommendations
- Placement: Below-article location is performing twice as well as sidebar (not what we expected)
Image attribution: WoCinTech Chat
What We’ve Learned
The art of personalization is, by nature, dynamic. As a result, we’re always discovering new benefits and honing our efforts to make our audience’s experience completely personal. That said, there are a few things we’ve learned so far about maximizing the efficacy of personalization:
1. Machine learning works. Our AI algorithms are out-performing random selections by far. And they continue to get smarter.
2. You need traffic. To begin creating anonymous profiles of visitors to track consumption habits, you need to drive traffic to your site. We estimate that you need at least 5,000 unique visitors per month.
3. Content feeds the engine. We estimate that you should have at least 50 pieces of content available and tagged on your site, and the ability to create at least 10 new pieces of content per month.
4. Measurement helps you learn. Machine learning takes continual monitoring and (human) learning, so you need access to the reports that measure performance. This insight will help you tweak the algorithms and eventually help shape your future content strategy.
Maybe it’s a little early to abandon your personas completely. But, when it comes to content, it’s time to get personal. For more information on Skyword Personalized Recommendations, reach out at firstname.lastname@example.org.
Featured image attribution: Pexels