Sometimes, losing a little control can be good, such as when you use cruise control to help your car get better mileage and help you avoid fatigue.
Similarly, automated personalization helps marketers increase conversions and revenue with little effort on their part, freeing time and energy for developing more creative campaigns. Automated personalization uses automated analysis of historical and real-time streams of visitor data to continuously determine what Web experience each visitor wants to have and then deliver it.
Overcoming the Fear of the Unknown
Drivers likely hesitated to use cruise control when auto manufacturers, on the heels of a fuel crisis, began adding it as an option on vehicles in 1974. What if the cruise control didn't turn off? What if it sped up on its own? Drivers feared losing control. Spurred by the promise of fuel efficiency with minor effort, however, many drivers began using cruise control in spite of their concerns. They realized that they actually had a lot of control; they decided when to use it, how fast to go, and when to turn it off.
With automated personalization, marketers likewise fear putting what they perceive as "blind trust" in a machine.
In reality, marketers can prime the machine with their knowledge of key visitor segments and their preferences. Marketers can specify success metrics and can turn campaigns on and off at will. Automated personalization does the heavy lifting of real-time decision-making and targeting based on a deluge of visitor data.
Although automated personalization promises huge wins like these, surprisingly few marketers use it.
Most marketers cite limits on budget, time, and skills as top barriers to automated personalization, according to our 2014 Digital Marketing Optimization Survey [PDF] results. However, the top 205 of surveyed organizations have a conversion rate—the rate at which marketers are turning prospects into customers—of 4.5% or more, at least 1.7 times the average (2.6%).
Let me nudge you to lose a little control with automated personalization by pointing out its strengths.
1. Automated personalization unlocks more value from your visitor data
You test one Website experience against another in A/B testing. Over time, the system determines which of the two experiences, on average, all visitors prefer, and serves that experience to them all.
Moreover, dividing your visitors into two basic groups can help you serve more personalized experiences. Consider what would happen if all online banking customers received the same offer for free interest-bearing savings account when they logged in—whether they had that account or not. Segmenting website visitors into those two groups lets you target each visitor with a far more relevant offer.
With automated personalization, that segmentation and targeting happens automatically, based on hundreds of parameters like search behavior, time of day, referring source, device, or even offline purchase data. It does all this with no need for you to manually identify segments. And each swipe, tap, and click immediately affects what the visitor experiences next.
When you consider the volume of online traffic that will occur during upcoming holiday seasons, for example, and the variety of devices from which that traffic will originate, you quickly understand the value that auto-discovery and targeting of potentially thousands of valuable segments offers.
2. It lets you use the knowledge and insights you already have
Automated personalization can harness your knowledge of your key audiences, letting you set up rules that specify the experiences for which these segments are eligible. It then uses machine learning to extract more value from those segments by fine-tuning its content targeting based on micro-segments within those higher-level segments.
Some automated personalization solutions automatically identify interest areas of a site—for example, world news, local news, and sports for an online news site. Automated personalization determines which interest area your visitors are most interested in and serves them the right experience accordingly.
3. It just gets better all the time
In A/B testing, when the test determines a winner, you typically stop running the test and serve all visitors the winning experience. That experience continues to be served until you initiate a change. An extreme example that highlights the potential problem with this approach is swimwear offers that won in July still being served in November. That might seem obvious, but the truth is that various factors that change over time (such as pricing, competing offers, and seasonality) affect visitors.
Automated personalization offers "always on" optimization of personalization. It continuously collects data about visitors and the performance of each offer and adjusts the machine-learning models to reflect the most current visitor behavior. This allows it to serve winning offers at any point in time rather than based on data collected over a limited, past testing period as A/B testing does.
4. It exploits all your data
Most companies have access to Web analytics, CRM, third-party demographic data, and call center data, or a subset of those data sources. Though that data can help paint an accurate picture of the visitor, the sheer volume of this data makes using it to follow visitors and make meaningful inferences about their propensities almost impossible.
Automated personalization can gather data from all those sources, and it uses machine-learning models to extract predictive information, which lets it make more informed decisions and ultimately serve visitors more relevant content.
Though A/B testing and rules-based targeting have their place in the marketer's toolkit, automated personalization may be a welcome addition to this kit, given its promise for increased conversion and revenue.
But don't expect to sit back and let automated personalization do all the work; your marketer's experience is still needed. Prime the machine with your knowledge of your site's visitors and their preferences, but then let automation take personalization the extra mile.
