After Omniture's acquisition of Offermatica late last year, I made a note to catch up with Chris Duskin, who is now part of the Test&Target team at Omniture. Testing, optimization, and targeting solutions have come a long way since my first experiences with Offermatica, Optimost, and others, so I wanted to tap Chris for ideas on how to apply these advanced technologies to the task of optimizing social commerce and social influence marketing. I've spoken with a number of customers and prospects about their plans to invest in this area, so I hope this interview is helpful to those efforts.
1. Let's discuss Omniture Test&Target – I associate Test with Offermatica and Target with TouchClarity, two companies and technologies that Omniture acquired separately. Why do they fit together so cohesively?
The capabilities that TouchClarity and Offermatica delivered separately didn’t really overlap and are even more powerful when used together. People often think of basic A/B testing as being on one end of the online optimization spectrum and fully automated one-to-one predictive targeting as being on the other. But in between are a variety of other capabilities like geo targeting, targeting to marketer-defined behavioral segments, multivariate testing, and automated champion-challenger tests. Marketers can benefit from all these capabilities and, with Test&Target, can use them together to truly change the game for their business. You may learn from a one-to-one campaign on your home page that a particular behavior is predictive. Well, you might then target the segment of visitors that hasthepredictive characteristic in a landing page campaign. At the same time, you might be running a multivariate test (MVT) on your registration page and an A/B test on calls to action throughout your site. Test&Target averages about 14 active tests or campaigns per customer at any time, which is a testament to the impact of putting marketers in control, enabling them to try ideas quickly, and automating key parts of the optimization process.
2. Our customers are often interested in A/B testing the impact of customer reviews and other user-generated content. What are the key considerations and test design factors to keep in mind when planning this type of test? What types of impact should customers look for, beyond the obvious like conversion?
Every test will have a specific goal and will be associated with a success metric. A company’s initial tests typically focus on conversion rate or revenue per visitor, but more sophisticated tests also consider metrics like registrations, cart abandonment rate, newsletter sign-ups, and participation in a site’s community. With product reviews, a key event is always review contribution. But whether people page through reviews as a result of your test is also important since it reflects meaningful engagement with the content.
In addition to thinking about success events, marketers should also consider visitor segments when creating their tests. Different types of visitors will respond differently to your alternatives. What matters to the person who frequently contributes to reviews might not matter at all to someone who only reads reviews, which of course suggests that I need to target different experiences and content to different types of visitors. We published an interesting case study from CNET related to this concept. On their mp3.com site, they relocated an artist photo widget to the top of the page for visitors who had previously interacted with that widget in its original, below-the-fold location. They saw a double-digit increase in engagement for those customers. And, the best part of the story is how quickly they turned a simple idea into an effective test using Test&Target. It took less than an hour to create the visitor segments, design the test, and launch it. Nobody from IT was even involved.
With product reviews, there are a lot of opportunities for testing and targeting beyond the product detail page. How can I use product review information in gift registries? Do star ratings in the cart reduce abandonments? Do I get more reviews when I target a homepage review solicitation to past purchasers?
3. The possibilities of behavioral targeting based on social engagement are virtually limitless – how should companies leverage the behavior of writing a review, asking a question about a product, or simply using reviews (sorting, searching, etc.)?
When people interact with UGC, they’re telling you something about what is relevant to them. Marketers can use these signals to target more and more relevant experiences. Of course, they’ll need to use testing to evaluate their efforts because people’s interests change – what was effective last year, last month, or last week might be less effective today. Companies using Test&Target to capitalize on this opportunity have developed a culture of optimization that embraces the ongoing and iterative nature of testing and targeting. Marketers are empowered to make site changes without waiting for IT, can try new ideas quickly, and can use predictive models that respond instantly to changes in interests. Gone are the days when you had to write one MRD to put up a new promotional banner and another MRD to pull it down a week later.
4. How are Test&Target customers using testing & targeting off site, beginning with an online advertisement, to set off a chain of dynamic experiences that result in a purchase on the site? Can you walk through a real scenario?
Your customers don’t interact with only a single page, and you don’t engage with them only on your site. Optimizing customer experiences across all interactions from the point of acquisition, through the conversion, and in your retention efforts is very powerful. We worked with Bantam Press a few months ago on a campaign to promote a book called SuperCrunchers. One test involved serving different display ads for the book on NYTimes.com and reinforcing those ads on-site after the click. When ads were reinforced on-site, results were dramatically higher than when the on-site experience was unrelated to the ad. And, after the best banner creative for click-through was determined, different types of on-site reinforcement were evaluated. We’re also seeing an increase in email testing and in testing combinations of email and on-site elements. For most companies, even a small increase in conversion off email can have a substantial impact. But testing email in the traditional, serial manner can be slow and inconclusive. With Test&Target, a single email template can be used to evaluate different content – one list of recipients, one email template, many email versions, and a clear winner delivered automatically at open time.
5. How can Test&Target be used to define and analyze different customer personas? For example, many of our clients have “super reviewers” that have written hundreds of reviews, but many consumers may write only a single review. There are various types in between. How can Test&Target help to identify those types?
The anonymous visitor profile in Test&Target is completely open to the marketer who can capture the behaviors associated with different personas. As a marketer, I may have special incentives or offers for people who have contributed over a hundred reviews. I can use the Test&Target profile to store the number of reviews written for each user and then target my incentive only to people who are in my “super user” segment. I might also tell the segment of users who have written 50-75 reviews about the benefits of becoming a “super reviewer”. Combining the Test&Target profile with SiteCatalyst or Omniture Discover, which gives you real-time insight into segments that matter, becomes very powerful. It allows you to easily and quickly act on your analytics and deliver relevance.