The digital age has brought us more and more data, as well as increased computational power to be able to process and analyse it. The newer and more popularised areas of Data Science tend to be in the realms of Artificial Intelligence and Big Data Analytics, but there are many types of Data Science that have been historically useful in business for much longer.
Statistical Data Science has been used by businesses for decades. For example, regression analysis helps us to understand relationships between variables, and can be used to forecast and make smarter decisions for pricing strategies, inventory management, understanding supply and demand, operations management and so on. What we hear less of, is its use towards the improvement of Customer Experience (CX); even though CX is recognised as the most important competitive battleground of business.
Speaking to business leaders from across industries, we have found that the top challenges with CX Transformation or Innovation are; (1) understanding where to focus change, (2) meeting customer and business goals, (3) getting buy-in from C-Suite decision makers, and (4) sticking to budget.
In light of these challenges, we at The BIO Agency have spent the last year building a framework to improve the success of CX projects; by combining our expertise in Data Science with Behavioural Science and Consumer Psychology. We call it, Experience Performance Modelling, or ‘XPM’ for short.
Our XPM framework helps customer experience projects by focusing ideation, and forecasting the impact of ideas on goal metrics.
This article gives an overview of the fundamentals of this framework, but please get in touch if you want to fully understand the unique way in which we use advanced statistics and psychology, or to see a demo of XPM in action.
In a world where resources are increasingly finite, we need to take a focused approach towards improving experiences, whilst aligning both business goals and customer needs.
What I mean is, we don’t want 100 okay-ish innovative ideas, we’d rather have 20 exceptional ones that we know are going to impact what matters most. Doing that will give us the most bang for our buck.
Step 1 – Identify Goals
So the obvious first step is to define what the ‘bang’ is – What is the goal and related metric we are trying to drive? Referencing your organisation’s Measurement Framework should give you the answer; common examples include customer satisfaction score, retention, Net Promoter Score, revenue, etc.
If you don’t have a Measurement Framework, we recommend building one (it’s one of the first things we do on many projects).
Step 2 – Hypothesise goal drivers using consumer psychology
We then need to hypothesise what is driving that goal metric – this is often done by using market analysis and customer research – but a useful way to theme goal drivers is with Consumer Psychology and the theory of Utilitarian vs Hedonic consumer needs.
- Utilitarian products and services are purchased for their practical, functional uses.
- Hedonic products and services are purchased to fulfill emotional and sensory needs – feelings of pleasure, fun, enjoyment.
In reality, a product or service is rarely just one or the other. Fewer people would buy a luxury watch (an item bought to inspire hedonic feelings of power or wealth) if it could not tell the time of day (its core utility); and many traditionally utilitarian services (e.g. banks, telecomms, etc.) are now trying to address hedonic needs to create more brand loyalty.
The point is, when creating hypotheses, we should consider goal drivers across both Utilitarian and Hedonic aspects of a service.
With a broad research question like, ‘What aspects of the online retail experience drive customer satisfaction scores?’, we can create hypothesis examples for each:
|Needs Category||Reseach Question||Hypothesis||Null Hypothesis|
|Utilitarian||Does order delivery time impact customer satisfaction scores?||Customers who have fast order delivery times will have high customer satisfaction scores.||There is no relationship between order delivery time and customer satisfaction scores.|
|Hedonic||Does the enjoyability of the shopping experience impact customer satisfaction scores?||Customer who enjoy the shopping experience will have high customer satisfaction scores.||There is no relationship between an enjoyable shopping experience and customer satisfaction scores.|
At this point, they are nothing more than hypothesese. We do not yet know that they affect customer satisfaction scores, and the relationship strength will certainly differ.
Step 3 – Collect and analyse data (the Data Science bit)
Once all of the above has been defined, we are in a position to collect data from customers using a quantitative survey with numerically based questions about our goal metric and hypothesised drivers.
Using a number of statistical modelling techniques (including advanced forms of regressions analysis) we can analyse survey responses to identify the strength of relationships, benchmark performance, understand driver groupings and compare customer segments.
Most importantly, knowing which aspects of a customer experience drive business goals allows us to confidently focus CX transformation efforts.
Using the previous example, even if both ‘order delivery time’ and ‘enjoyable shopping experiences’ are correlated, if we have finite resources then we would focus on the one that’s most correlated.
Now we can ideate new business opportunities with the confidence that we’re focusing effort in the right place. I’d call that a very useful application of Data Science!
But wait. Once we have addressed the first challenge and ideated in our focus area(s), we still won’t know the real impact of those ideas on our goal metric. Will our ideas increase performance by 3%? 10%? 42.57483%?
This becomes our second challenge. If solved, we can create a very strong case for change.
Unfortunately, there is no one size fits all approach, and we need to consider a number of things to decide on the best forecasting method. Having said that, below are some worthy examples of how we have previously tackled the problem.
Option 1 - BIO’s Statistical Models
Statistical Data Science can once again help us when we are dealing with particularly innovative ideas. We, at BIO, have developed a statistical model which uses the data collected when focusing ideation (discussed above) with further research and a set of assumptions to forecast the impact of different combinations of ideas.
- Best used when forecasting customer satisfaction metrics (e.g. customer satisfaction scores, Net Promoter Score, retention).
Option 2 - Business Casing
A traditional and effective method is to use your company’s financial data and use robust assumptions from research to develop a detailed financial appraisal.
- Best used when forecasting financial metrics (e.g. revenue, net present value, internal rate of return, return on investment, payback period).
Option 3 - User Testing
Building a Proof of Concept or Minimum Viable Product in combination with user testing techniques (e.g. proposition testing, A/B testing, multivariate testing) can become very scientific if done correctly. Results can be used extrapolate results and understand impact.
- These can be used on both customer satisfaction metrics and financial metrics.
Option 4 - Historical Data
Of course, if there are applicable data sets from similar transformations, then analysing and forecasting with that data is another great way to understand impact.
- These can also be used on both customer satisfaction metrics and financial metrics.
Regardless of which method is most appropriate, what you end up with is a number of innovative ideas and the knowledge of what the real impact will be on the customer experience and business goals.
To finish things off, the forecasts can be combined with other criteria to prioritise, group and roadmap; giving you everything you need to confidently plan impactful change.
We think there is great power in introducing this kind of methodology to focus your resources, forecast real impact, and make data-driven decisions to achieve greater success.
Make customer experience decisions with Data Science now
Today, more than ever, external forces are pushing businesses to their limits. To survive, organisations need to adapt services for progressively higher customer expectations, but with increasingly finite resources.
Don’t let your transformation projects fail. Find focus and understand the impact of projects before you invest.
Talk to us if you want to know more about how you can confidently make impactful change, or to discuss our Experience Performance Modelling (XPM) framework in more detail.