Here at Balihoo I recently developed a content or resource recommendation system that allows our website to dynamically feed recommendations to every visitor of our site. When setting up automated coding to nurture our current leads, we realized we had far too many content pieces, and we were guiding users down too many different paths. Instead of hard-coding in set “paths” of content where leads would be fed a particular piece of content, we wanted demographic and individual lead behavior on our site to drive their personalized recommendations.

recommendation
Is content recommendation right for you?

Personalized product or resource recommendation is extremely useful when customers do not fit into a few set buying personas. Count how many types of prospects you have. If you can only count a handful, then investing time and money in a dynamic content recommendation system is probably not the best idea. Simply use tools to identify which of the types those customers are, then push the pieces of content that are relevant to them with static nurture cycles.

Content recommendation, however, is extremely powerful if they don’t fit the above model. The basics of a content recommendation system involve scoring each customer against every single piece of content that your company offers, and feeding them the highest scoring piece (or pieces) at any time.

Setting up content recommendation

To be able to set up an effective content recommendation system, consider the following.

First, how is the content you serve related? Bring together the product team, marketing team and possibly sales, and figure out relations between pieces, and come up with a scoring system for relation. For example, we know that if a customer downloads a particular white paper, then it is very likely they are also going to be interested in a related webinar and info graphic, so we create scoring correlations between these items. Answer the question, “If someone is interested in this, what are all the other things that might tell us?” for each piece of content, and make sure to score the pieces as related.recommendation_ss

Second, use known demographic information about leads to score content against a customer. There are many tools available that will use IP address information and allow you to determine an industry, company size and location, and more before a lead supplies you with any information. Use this information to assess all your content pieces. From each piece of demographic data you can deduce from a lead, ask “What does this tell us they might be interested in?” Set up scoring rules for each piece of content that take that into account.

Lastly, don’t neglect other behavior-related information you may have about a lead. Things like the referring URL that the lead came from can tell you information about them. Did they enter from your YouTube channel? They might enjoy videos and webinars a bit more. Have they been on your website multiple times over a few years? Maybe the more in-depth, research-driven content is what they are looking for. On the other end of the spectrum, did they first encounter your site and spend an hour and a half looking through quite a lot of pages? Score them high for a live demo and try to push them to request a sales call.

Using Scoring

Once you have the scoring and recommendation system set up, the fun begins! At Balihoo we use the content recommendation in a few primary ways. Our main use (and original reason we started considering using a content recommendation system) was to drive dynamic “You might like this…” links at the bottom of each download page and at various places around our website. How we have it implemented currently, we offer the piece of content that is ranked the highest for a given lead. So, every person that visits the exact same “Thanks for downloading…” page will see a “You might like this” link that is unique to them, based off of their behavior and known demographics.

recommendation_ss2
Secondly, we send out emails to leads suggesting that they download or view the content we know they might be interested in. While various activities could be used to trigger these emails, the one we primarily use is follow up to downloads. Other actions could trigger these emails as well, or simply use the content recommendation as a part of other regularly scheduled email correspondence with a lead. For example, to the bottom of a normally static newsletter you could add content recommendation that would essentially create a customized email experience for every lead that receives your newsletter.

The possibilities really are endless. Once your CRM system has the “knowledge” of what a lead is interested in, you can use this to speed up your buying cycle, increase your conversion rates, fill your pipeline and pass much more qualified leads along to sales. Think outside the box with this! How about when you alert your sales team to a new MQL, in the name notification you let them know a few pieces of content that the lead may be interested in that they haven’t downloaded yet. That would give your sales agent a great way to start the conversation!

Results and moving forward

While we have only been using content recommendation for just over four months, we have tripled the number of email interactions with a customer after they download content, achieved a higher open rate (not to mention a 20-25% clickthrough rate), and sped up the buying cycle by leads that have engaged with our content recommendation by 13%.

Related posts

Leave a Comment