Why measuring CX ROI is being undermined by meaningless means

Some time ago (OK, a long time ago), I wrote an article measuring and comparing the return on investment of different CX strategies in which I illustrated that the relationship between customer satisfaction and the 'effort' required to achieve any absolute improvements is rarely linear. In other words, it is usually a lot harder to move a customer from neutral to satisfied than it is to move a dissatisfied customer to neutral. I also talked about how the effort (cost) curve is different for different kinds of channels and experiences.

In this article, I am going to take this a bit further; converting 'satisfaction' into 'receptivity' (i.e. the degree to which a customer is likely to listen to you and act in a way that impacts the bottom line). Put another way, how a combination of circumstances and the way customers feel about you directly influences their propensity to spend money with your brand.

I am sure we have all seen diagrams something like the one above; the implication being that the more satisfied a customer is, the higher the Average Revenue Per User (ARPU). If you can increase the satisfaction level, you should expect higher revenues. 

Ignoring for a moment the fact that it is rarely a linear relationship, this graph conceals a lot of important details: for example, it doesn't discuss the cost of achieving the higher level of satisfaction (which you need to calculate an ROI), neither does it break down the customer population into different segments, and it definitely does nothing to explain the causal link between the two axes; what are the drivers of satisfaction and how does that equate to revenue? 

The meaningless of means

I hate 'averages' (means); there is no such thing as an average customer, or an average employee; an average without context tells us very little. For example, how distributed is the population? Is everyone close to the mean, or more widely spread out? 

The mean of 0 and 100 is 50, and so is the mean of 45 and 55, yet I would suggest that these represent two very different sets. 

Look beyond an average measure of satisfaction - what is the 'shape' of the data?

If your satisfaction scores are widely dispersed, you have a problem with consistency, and a problem with consistency translates into 'unpredictability'. As human beings, especially in uncertain times, we want brands to be predictable, even if they are not great. There is an expression I like that covers this very well: "Better the devil we know". In other words, customers would rather accept a predictably bad experience, than take the risk it might get worse.

Before you try to improve mean satisfaction, get consistent!

How customers feel matters

But there is something else the graph above doesn't show: different customers, with different feelings and attitudes about your brand will behave and respond differently to your actions.

Rather than go into a lot of theory about the influence of emotions on customer behaviour, I will state an assumption: customers who feel kindly towards you are more receptive (prepared to listen to you) than those that don't (a gross over-simplification, but bear with me).

In the graph above, I have given three very simple Empagraphic Segments (check out Anthrolytics.io to find out what that means). I have given them descriptive names, but that's only for the sake of convenience.

The 'Overlooked and Neglected' are often existing customers who, whilst not actively thinking of defecting, don't respond to sales or marketing campaigns either. However, if they are made to feel a little more valued, become a bit more receptive and then will start spending more (in one population I analysed, these offered the best ROI in respect of overall customer life time value).

The 'Surprised and Delighted' are customers who are already very receptive and can represent the highest value segment in terms of net revenue. These customers love your brand and are open to new offers and experiences - show them other ways you can help them improve their lives, and they will ask where to sign.

The 'Neutrally Satisfied' represent a group that feel you are doing an 'okay' job of meeting their needs and may place you amongst a group of brands offering similar products. This group can be troubling, because, whilst moderately happy, that's not what's important to them; you can make them happier, but they don't spend any more money as a result. Blindly marketing to this group could result in a negative RoI, because the campaign can cost more than the increase in revenue.

Let's take one final group, the 'Deeply Disappointed' (not on the graph) - these are current customers who have had predominantly negative experiences with your brand. To be frank, it's probably unlikely you can do anything to recover these relationships, and spending money trying to retain them is extremely unlikely to result in a positive ROI.

In conclusion

Simple lines that show a correlation between two axes never tell the whole story. And neither does thinking of all customers as if they are the same, and especially that they all feel the same way about you as their relationship with your brand unfolds.

And I have left a lot out too; what about the economic imperative? For example, affordability, or availability (are you the only game in town?). Customer behaviour can be complex, but if you want to really know why customers (or employees) do what they do, and what they are going to do next, you are going to have to start thinking about them in a different way.

We live in a world that is Volatile, Uncertain, Complex and Ambiguous (VUCA) and businesses are facing a lot of economic headwinds as well. CX professionals are going to have to do more to explain the economic mechanisms behind their ROI calculations. We have been using segmentation in sales and marketing for decades, based on quantitative criteria, isn't it about time we started to do the same at a qualitative level? If you want to find out how, drop me a line.