- Customer experience is vital for company’s future success, how you communicate with your audience must be in a simple and straightforward manner.
- There are two ways to segment customers; either by using opinion based insights by drawing input from several people involved with the company or by utilising machine learning which mathematically identifies groups that exist within a data set that are reflective of the customer base.
- To accurately realise customer segments, leverage the established insights of opinion-based personas with impartial, fact-based clusters.
Historically, marketing communications comprised largely generic messaging. Low levels of personalisation were employed, if they were at all, to distribute a message to as many people as efficiently as possible. Low cost tools made it simple and easy to do this. This is an example of failing to consider digital transformation properly. Simple replacement of an analog process with a low-cost digital one was too hard to resist.
This approach is no longer effective. Businesses and consumers have reject
ed generic messaging and frequently unsubscribe from these communications. Data driven, highly-personalised content experiences are the new standard. Failure to adopt this means the market is more likely to disengage with your brand (CMO 2015). More recent studies also conclude that the need for personalised interactions continues (HubSpot 2016).
The conclusion from this trend is that delightful customer experience (CX) is vital to future success. If we cannot communicate with our market in a simple, straightforward manner with content relevant to the recipient, we are unlikely to provide a satisfactory customer experience. However, to achieve this we need to seriously understand our customers and the journeys they follow.
In this blog we will demonstrate methods for understanding customer segments that are both traditional and contemporary. Further, we will show the substantial improvements possible when these methods are combined.
Thinking About Customer Segmentation
Customer segmentation is the practice of organising customers into discrete groups with similar characteristics. This is typically done in one of two ways: Personas and Clustering, which are otherwise known as opinion-based and fact-based respectively.
Personas: Opinion Based
A persona is a conceptual model of a person with a name, characteristics and a specific way of doing things; it is a fictitious or imaginary person. From a marketing point of view, the aim of a persona is to be able to see your products and services from their perspective.
Persona development is a collaborative venture drawing on input from several people in the organisation that “understand” or “know” the customer. This can involve people from marketing, sales and customer service. The thoughts and experiences of the contributors determines the personas chosen; the more diverse their perspectives, the better the outcome.
There are several drawbacks to developing personas:
- Contributors with many similar perspectives can lead to skew or bias;
- Subjective views and a shallow understanding will directly impact the quality of the outcome;
- Changes in customer behaviour can take time to register. It is therefore difficult to maintain an up to date view due to this latency; and
- The process is not fact based. Opinions and assumptions drive the responses.
Opinion based persona development is a tried and tested approach. Despite the drawbacks it is nonetheless a good starting point. A strong focus on the quality and balance of contributors is key to the value of the outcome.
Clusters: Fact Based
An alternative method of understanding customers is cluster analysis. This leverages Machine Learning, a sub domain of Artificial Intelligence (AI). This method mathematically identifies groups that exist within a data set that are reflective of the customer base. This fact-based approach immediately limits the errors and limitations in the opinion-based or human approach. It can segment customers over many attribute dimensions and create homogenous groups.
Just as the opinion-based person approach depends on the reliability of the opinions and assumptions, this approach is only as good as the quality of the data used; it is truly garbage in, garbage out. There are also limitations in the application of machine learning. It is not a common approach, the quality and experience of the person interpreting this will impact the result.
On the face of it the fact-based method will provide a more reliable outcome. If we wish to avoid the limitations of human interference, a fact-based approach is surely better? However, the opinion-based approach also has merits and is the current go-to approach for creating personas.
Below we contrast the two approaches:
Two is Sometimes Better Than One
Despite the strong differences between an opinion and a fact based approach, one does not preclude the other; the approaches are complimentary. To accurately realise customer segments, leverage the established insights of opinion-based personas with impartial, fact-based clusters. The approach taken depends on whether a defined set of personas exists:
- Validation Method: This is applied when personas already exist. These personas are then validated with results from a cluster analysis. The validation method provides an unbiased perspective to confirm persona hypotheses.
- Generation Method: This is applied when no personas exist. Cluster analysis is used to gain an understanding of customers, which is then augmented using personas. In this way a much improved customer understanding is achieved.
Optimal customer experience (CX) is key to customer success. Through a deeper, richer understanding of customers we can deliver more personalised communications and improve brand engagement. This improved understanding of customers is achieved by combining traditional, opinion-based personas with fact-based clusters.
If you are interested in receiving a customised Customer Clustering Report for FREE, please click here!
Alternatively, if you want to find out more in regards to how AI can help your company, please contact us at: +61 2 8599–3114.