Gotta Catch Em’ All! — High-Value Customers not Pokémon!
by Andrew Koperwas
posted on 08-26-2016
How well do you know your high-value customers? How are you using analytics to drive more effective marketing practices at your company? How can I catch a Mewtwo in Pokémon Go? This blog post will help with some tips on the first two (but contact me if you know someone who can help with the third).
To effectively prioritize your marketing dollars and time, it’s important to have an in-depth view of your audience base and how they behave across your properties. Luckily, leading analytics tools, such as Adobe Analytics, can help. This subset of customer analytics or customer-journey analytics allows marketers to identify key behaviors, traits, and actions that are exhibited by prospects or customers, leading to the creation of much more powerful segments. Therefore, marketers are able to realize improved ROI, decreased costs, and increased satisfaction.
While I won’t go into Pokémon Go tactics here (make sure to save up your Pidgeys for the Lucky Egg!), I will force an analogy between the game and some of the features within Adobe Analytics that will help your organization see immediate insights that can drive measurable results.
Who Are My High-Value Customers?
First, what do high-value customers mean to your business? For a few, customer lifetime value (CLV) is a common metric used in their organizations, and they have clear understandings regarding customers’ values. Pokémon Go makes it easy, with each Pokémon having a different power and level. Unfortunately, customer value isn’t so simply determined, and for most organizations, CLV doesn’t exist or isn’t easily used. For those in retail, this can be more straightforward, as you can use calculated metrics to create a good proxy for this. For others, such as those in media, this may be more difficult. You may want to create some sort of engagement score from calculated metrics based on your goals (number of articles read, time onsite, return visits, email signups, etc.).
What Are the Behaviors of My High-Value Customers?
While creating calculated metrics seems easy enough, what if you don’t always know what these behaviors should be? For example, if you’re a product manager in charge of brokerage accounts for Goliath National Bank, your best tactic for gaining new accounts is to upsell current checking account holders. Instead of just blanketing all current checking clients, let’s understand all our current customers that became brokerage clients after being checking clients. By filtering this segment, we can then use a tool like pathing to see the path someone took before becoming a customer. In data workbench, we can use the latency table to view the events and days before a conversion occurred to analyze specific online and offline touchpoints that our clients may have interacted with before converting. We can also see post conversion to understand whether we have the right content to guide them to better experiences.
All of this gives us much better views into what some of the key triggers are before someone converts, but what if we want to see the differences between segments? For instance, I have a few Pokémon that have the same combat power, but I only have so many resources to strengthen them. So, how do I choose which Pokémon to train and make my main one? In real life, how can we understand the differences between our segments?
Segment Comparison, a feature within Segment IQ in Analysis Workspace, intelligently discovers the differences between your target audience segments through automated analysis of all your metrics and dimensions. Now, we can easily drop in “checking and brokerage clients” and “checking and no brokerage” or “high-value customers” and “low-value customers” and receive an interactive report that lets us see what the most important differences are between each segment.
Using these insights, we can identify the most statistically significant differences between different segments to drive better segment creation. Additionally, we can see the overlap between segments. For customers of Adobe Analytics and Adobe Audience Manager (AAM), all of your AAM segments are now available in each solution in real time. This allows you to use second- and third-party data, alongside features such as lookalike modeling, to expand your audience reach and see how these new segments overlap your current segments; it also enables you to identify overlap, avoiding duplicative or confusing marketing for better user experiences and reduced marketing spend.
Another simple tool for audience and segment exploration — also provided by Adobe Analytics Premium — is the audience clustering feature of data workbench. Most people think of audience clustering as segmenting on steroids, but you could also start this process with clustering and see where the algorithms identify statistically significant metrics or dimensions.
Select your input variables (such as those sourced from segment comparison) — the number of clusters, target population, and desired algorithm — and the feature will automatically analyze your data, dynamically categorizing visitors and generating cluster sets based on selected data inputs, thus, identifying groups with similar interests and behaviors for customer analysis and targeting.
For additional insights, this allows you to further explore what makes each cluster unique and gives you an easy way to take action and test your marketing campaigns on each cluster/segment. Using the tools mentioned before, you can understand where there is already overlap, bring the clusters into Audience Manager to expand and activate the segments, and then use Analysis Workspace to track and compare how these new clusters compare to your new segments.
I Don’t Like Pokémon Go, so I Just Scrolled to the Bottom for the Takeaways.
First, I saw eight people playing Pokémon Go this morning on my 20-minute bus ride to work, so at the very least, you may have picked up some vocabulary for the next time a gang of Poképlayers envelopes you on the street. Second, we went over a number of different features within Adobe Analytics that can help you gain insights on your different segments. Here are two in particular you should start using:
This is one of the most rapidly adopted features as well as one of the easiest to use. All you do is drag and drop two segments, and Adobe Analytics does the rest. See our YouTube channel for more on Analysis Workspace and Segment Comparison.
For Adobe Analytics Premium customers, audience clustering allows you to jump ahead a few steps. You can use clustering to find key differences in your data. See how to set up and use clustering here.