Web Engagement: Should Your Recommendation Engine Work with Real-Time Data?

By Arjen van den Akker (@arjenvdakker) Oct 20, 2011

Under the umbrella of “customer engagement” and “web engagement,” companies look to implement technology and processes to ride the next wave of the World Wide Web. After Web 2.0, companies now need to engage with their audiences.

And of course, it’s clear that visitors can no longer be treated as anonymous “guests” on a website, especially not when they have identified themselves by logging in. They should be recognized and serviced appropriately, with relevant information. And ideally, that experience should also be context sensitive, especially when someone uses a mobile device. But how do you make the content relevant?

When looking at B2B and B2C situations, it is important to consider the role of data analysis and recommendation engines, as well as asking the question of real-time versus batch, as the choice you’re making about that might have far-reaching consequences.

B2B vs. B2C: Aren’t We All Individuals?

There are many similarities between B2B and B2C engagement. In the past, B2C marketing was mainly about broadcasting, whereas B2B marketing used to be more direct and personal. But with the advent of the empowered and social consumer, consumers have become addressable as individuals as well. They are no longer anonymous faces in the crowd. Instead, they have a voice and a face. As a result, companies need to look for ways to adopt “mass personalization” in their marketing strategies. Apart from the fact that B2C personalization needs to deal with much larger volumes of people than B2B, there seemingly isn’t a huge difference between the two anymore. Or is there?

I Came to Your B2B Website, Now What?

In a B2B scenario, website visitors that identify themselves are likely to be stored in your CRM system or other back-end systems already, or they are candidates to end up there. Note that even though they are individuals, they are representatives of another company — hence we call it B2B. Because the number of companies that a B2B business deals with is limited, we are talking relatively low numbers of visitors here.

In many cases we see B2B companies linking up their website with their back-end systems to achieve the following:

Capture leads (through web forms, for instance) and store them in a CRM system
Auto populate forms with data from the CRM or other systems
Update personal profile details (self service)
Check membership to grant or deny access to certain areas of the site
Personalize web content based on known information

The personalization of web content is often executed by establishing rules that are executed against someone’s profile and other known data, driving certain real-time decisions about which content to show on the site. Because the volumes are low and the amount of personalized content is usually limited, the desired real-time approach is usually very achievable.

What about B2C?

How different is this from a B2C scenario? Broadly speaking, the B2B approach still makes sense, but in addition we are talking many more potential website visitors. And companies will have a much bigger desire to promote specific products. A typical online retailer will have thousands of products, and in some way they will need to decide which products to showcase to the visitor during a visit in order to upsell or cross sell.

Of course, we can apply the visitor’s own explicit profile and historical behavioral data to decide which content to show, as well as implicitly generated data such as browsing behavior, time of day, the device used to connect to the site, place of origin and the local weather there, etc. But because we’re talking larger volumes of visitors, we suddenly have a unique opportunity to also apply learnings from how other consumers with a similar profile have behaved in the past.

Another Realm…

This suddenly takes us into a completely different realm of customer intelligence (CI) professionals. CI pros are more experienced than their interactive marketing counterparts in most data-related activities. They know how to use a number of customer data streams, such as transactional, customer feedback, insight from statistical models and offline segmentation data to support campaign planning. They excel at building customer and marketing databases for direct marketing use. This often involves activities such as data integrity, data hygiene, standardization and business rules design. And while interactive marketers typically rely on digital campaign performance analysis, CI professionals focus on broader marketing performance and customer growth initiatives.

They should now lend their services to apply the same techniques to mine for insights into data generated across all channels, including digital. This will allow interactive marketers to accurately apply richer data-driven insights in content optimization decisions and ultimately personalize web content.