In today’s edition of Analytics in Action, let’s talk about how CLV (Customer Lifetime Value) can explode your ROI. First, let’s talk about segmentation in general and CLV as a specific segment you should bring into focus.
OK, so maybe you’re saying “What a waste of time. I already segment my market. “
I’m sure you do segment your market and choose the target market or markets more likely to purchase your brand. But, if you’re not using higher-level tools like CLV and predictive analytics to choose the RIGHT target markets, you’re likely leaving a lot of money on the table.
Customer lifetime value is just what the words say: projecting the total value of a customer across the lifetime of that customer – commonly the retention of the customer, rather than their physical lifespan.
Put simply, CLV is calculated using the following formula:
CLV = Revenue from a customer (over time) – Costs to acquire and retain customer
In calculating the revenue you consider the average order value (AOV) and how long the customer stays with the brand. Costs include incentives, marketing communication costs and other things incurred to get and keep your customer.
This formula sounds super simple. So, why doesn’t everyone use it?
Because things are often not as simple as they appear. For instance, consider this CLV calculation courtesy of The Darden School: Challenges in calculating CLV
In fact, Tristan Handy at RJMetrics says this about calculating CLV:
“We’ve found that calculating customer lifetime value is one of the single biggest challenges digital marketers face. Companies tell us that they spend countless hours with SQL queries and spreadsheets or pay thousands of dollars to consultants.”
Never fear. You don’t need to spend hours and thousands because they offer a free CLV calculator using information such as acquisition cost, retention rate, cost of capital, etc.
In a minute, I want to talk about how you can use calculations of CLV to explode your ROI, but first, I want to take a minute to discuss predictive analytics. That way, I can show you how CLV helps your business explode ROI both with and without using predictive analytics.
Here’s what InformationWeek said about predictive analytics:
“Predictive analytics is the next step up in data reduction. It utilizes a variety of statistical, modeling, data mining, and machine learning techniques to study recent and historical data, thereby allowing analysts to make predictions about the future.”
Unlike descriptive analytics, like Google Analytics, that only tell you what happened, predictive analytics try to predict what MIGHT happen in the future. Unlike descriptive analytics that dominates social media analytics, predictive analytics aren’t forecasts. They don’t show you absolutes, only predictions what might happen.
Now, you might ask, “Why would I want predictions if they’re not 100% accurate?”
Simple. Because they’re more accurate than expecting what’s happening now to continue, which is the backbone of most forecasting models.
Maybe an example will help.
Let’s say your #visits looks like this:
Notice the dashed line forecasts future visits based on a simple extrapolation where we expect historical growth continues. Likely, if you’re like most social media marketers, you make forecasts like this.
Predictions superior to forecasts
Forecasts easily go awry when unforeseen circumstances occur. For instance, imagine someone influential Tweets a link to your content?
I’ll tell you what happens to me when someone like Mari Smith or Jeremiah Owyang Tweet my links – my # of visits goes up by nearly 30%, well over projections. The increase extends over several days, making my monthly visit numbers out of whack with forecasts.
Now, maybe poorly forecasting visits doesn’t mean much (or it might also take your site down when your host can’t handle the traffic), but # of leads likely has a significant impact on staffing and budgeting. A significant increase in leads might overwhelm your staff, causing your conversion to drop and pissing off potential customers. Too few leads and your staff sit around costing you money. Poor lead forecasting throws off budgeting.
Using predictive analytics
Now that you have some appreciation for predictive analytics, let’s talk about how you do predictive analytics. For more insights on predictive analytics, check out this post on using predictive analytics to improve your ROI.
Commonly you build a predictive model; often a simple regression model using historical data. If you’re trying to predict the number of leads over time, you might build a model based on how the number of leads responds to things you control, like content marketing efforts, social media engagement, paid media, and things you don’t control, including economic factors like GDP or consumer confidence.
For instance, this image from Conversionxl:
The model lists potential factors figuring into CLV that help build a predictive model. Gather metrics about these factors, then build a simple regression model showing how much each factor contributed to past CLV. Now, use the model to predict what your future CLV looks like by plugging in likely values for these factors.
Data mining is a common tool for building predictive models based on “big data”. While building complex models might be beyond the skills of some small business marketers, it’s easy to hire a consultant to build the model. Then all you need is to plug in new data every period to get new predictions. Trust me, results easily repay the costs with huge dividends.
Using CLV with predictive analytics
I already showed how you can use predictive analytics to improve CLV. Now, let’s talk about using predictive analytics in combination with CLV to explode ROI.
Combining predictive analytics and CLV can show you which consumer segments to target based on either their spending (AOV) or retention. See, we finally got around to the topic that opened this post!
Again, Conversionxl shows the response of different segments based on CLV:
See how different groups of consumers show different loyalty. By focusing your social media on the group that’s least loyal you might retain more of them.
For example, you might offer members of the sad group something (money, free product) to get them to stay longer or you might focus efforts on increasing spending of the awesome group, offering prepayment discounts or other incentives.
Using predictive analytics to identify characteristic shared by members of each group also allows you to predict which prospective customers to target. For instance, you might find the awesome group shares certain demographic or geographic characteristics or interests – using the enhanced analytics provided through the Google Analytics Universal tracking code, for instance.
Knowing what awesome customers look like, you can target them when setting up Facebook, PPC, and other forms of digital advertising.
According to Custora, the margin of error for this combined CLV/ predictive analytics approach is negligible while the margin of error for traditional projections of CLV is 250%.
That’s why combining CLV and predictive analytics can explode ROI for your brand!
I realize this post was pretty dense and the concepts unfamiliar to some of you. I’m happy to answer any questions you have. Just ask in the comments below.