Measuring return on CX investment encourages management buy-in and creates a business case for future investment. But how do you do it?
In order to stay competitive and achieve a sustainable differentiation, companies cannot only focus on the success of a product or service. They need to invest in customer experience (CX).
Customer retention and increased revenue are results of positive experiences at various touchpoints throughout the customer journey. But the question is: how do you measure the impact and success of these investments?
In order to assess the effectiveness of CX management, it is necessary to identify and quantify business results. This data can then be used to build the business case for CX spend, and achieve management buy-in based on the identified return on CX investments (ROCXI), and create a business case for future developments and investments.
What impact does CX have on business results?
Often customer feedback is used to measure the effectiveness of CX activities. Although overall satisfaction and willingness to recommend are well suited for target setting and for measuring contact point performance over time, they are only values that reflect customer perception.
However, the ultimate benefit of CX can be measured by the extent to which positive customer experiences lead to better business results. In order to locate this evidence, it is necessary to bridge the gap between customer feedback and their actual behaviour.
The solution to this task lies in the existing data pool and the intelligent handling of it. With the right data linkage, the effectiveness of CX activities can be proven. It explains the relationship between the customer experience and downstream business results such as customer retention, revenue and profit.
How do you ask the right questions?
Linking different data sets is known as linkage analysis. In most cases, the relationship between a CX ratio and monetary value is of foremost interest. So the main question to be answered could be: "How much more revenue will I earn, if the CX index, such as the Net Promoter Score is increased by one percentage point?" Here the CX-KPI is the effect on the business result.
An alternative question could be “What level must the CX index be raised to, if the ultimate goal is to reduce the churn rate by 5%?” The result of the analysis then allows us to define a target value derived directly from the targeted business result, rather than a generalised target: "We want to improve our NPS by 10 percentage points over the next 12 months.”
Simple linkage analyses only require two variables, such as the relationship between satisfaction and turnover. More complex analyses include operational data - especially with regard to employees or processes.
Linkage analyses can be used to solve a number of other questions that are crucial for the design of a CX programme. For example, which CX ratio is best suited as a reliable indicator for a company's economic target figure? Does the internally developed CX index provide a better control of success than a standard Net Promoter Score?
In addition to these conceptual questions, linkage analyses can also be applied for operational tasks. For example, which internal processes need to be adjusted to improve the customer experience as effectively as possible? Or, which customer segments should be prioritised in order to achieve the CX goals within a set timeframe possible?
Simple linkage analyses only require two variables, such as the relationship between satisfaction and turnover. More complex analyses include operational data - especially with regard to employees (fluctuation, training, retention) or processes (time to market, resources, disruptions). In this way, forecast models can be created that are not only more relevant to action, but are also suitable for simulating the implications of internal changes on a CX strategy in advance.
Willingness to recommend
For example, a bank's new customer analysis examined a customer's value over a period of 24 months, based on the customer's willingness to recommend a new customer. While there was no significant difference between promoters and passives, the customer value of a promoter was 6% higher than that of a detractor.
Promoters and passives were also on par with each other in terms of dismissal, whereas detractors were more likely to migrate to a competitor with an 11% higher probability.
In a further step, the bank was also able to calculate that a conversion of 1% neutrals to promoters could generate additional income of just under 300,000 Euros. With a development of 1% detractors to promoters, however, it could realise an additional EUR 5.5 million.
The analysis thus not only established the link between satisfaction and revenue, but also made it clear in which customer segment the bank should increase its focus from an economic standpoint.
Five key steps for successful linkage analyses
No matter what type of linkage analysis you choose, the process always follows the same steps and can be divided into three phases: data audit, exploratory research and predictive modeling/projection.
The first step is the data audit, in which the aim is to review the existing data. What information is available? How is the quality of data? Who owns the data in the company and can it be used? And of course, which values are suitable as a link between the individual data sets? As best practice the link is made at individual customer level. Where this is not possible, aggregated data can usually be accessed.
No matter what type of linkage analysis you choose, the process always follows the same steps: data audit, exploratory research, and predictive modeling/projection.
The connection is then formed by individual customer groups, specific periods of time, regional clusters or business units. In these cases, standardisation of the data is often necessary. For example, larger car dealers should also achieve higher sales, simply because of the higher number of employees and not necessarily because of better service.
During the exploratory research phase we learn about the data and the patterns between the data sources. This phase often reveals valuable relationships within the data that organisations were not aware of or did not have a proof about before. Pedictive modeling, or at least some form of a projection of a simulated change in CX and its impact on customer behaviour, is often at the heart of the analysis.
In principle, many of the methods commonly used in marketing come into consideration for the actual analysis - from purely descriptive approaches to advanced multivariate analyses. The selected method must not only reflect the variety of data applied but also to a large extent the type of data and the expected connections.
Therefore, in order to carry out successful linkage analyses, CX managers need to follow these five steps:
- Clarify the objective of the analysis and define what the results will be used for;
- Start with simple analyses before adding variables to a model;
- Proceed step-by-step and plan enough time for viewing and checking the data at the earliest possible stage;
- Consider aggregated data if it is not possible to link at individual customer level;
- Use the validated relationships to strengthen the relevance of CX to the success of your company.
Oliver Skeide is diretor of expert services EMEA at MaritzCX.