User Personas: How to Create Them - Part 3 - Review

 

In part 3 of the topic, we will try and complete the discussion of User Personas.

As like the earlier articles on this topic, this article too is inspired from the learning at CXL Institute.

We will explore the identification of clusters and then move to discuss how to build the archetypes of user personas.

Identifying Clusters

As we were discussing in the earlier article, it is beneficial to use the computer software to identify the clusters.

We can have different visualization models to depict them.

We can also average the cluster members to find the average cluster member score.

Though the individuals vary in the responses to the factors and scores, usually they do vary around the mean or the median.

Hence, we can find the mean and average the scores within the clusters, and build graphs for each type of persona.

The graphs represent the average member from each cluster.

With the graphs, it becomes clearer how the clusters differ from one another.

There are a lot of tools and applications available to do the cluster analysis.

If you already have an analyst in your team, that person can be of benefit.

You can discuss with that person the tool to be used.

One of the tools suggested by CXL Institute is R. R is a statistical programming language.

It has a very easy way of data analysis. One can put in the data and require a particular cluster analysis from it.

It also shows the results of the analysis.

One can easily learn the programming language by oneself or with the help of someone who know the language/

Other tools suggested by CXL Institute are GNU PSPP, IBM SPSS, and wessa.net.

Stereotypes to avoid

If you are not handing the data to the analyst, then there are a few stereotypes in cluster analysis that you need to avoid.

One thing we need to do before we put the data into the cluster analysis is to ensure that the scores are normalized. In other words, to ensure that they span across similar ranges.

Another thing to check are the outliers and how much influence each outlier has on the clustering solution you are trying to find.

The third thing which is recommended by CXL Institute is to visualize the data. It becomes easier to visualize if the factors chosen are two to three.

Basically, what the visualization helps in is identification of individuals that are outstanding from the cluster that you are analyzing.

This helps in identifying the need of another cluster to cater that one individual along with other members of the existing cluster that do not really match up with the existing clusters or have few characteristics similar to the new cluster formed.

Another way of clustering is called the hierarchical clustering.

This particular type of clustering pulls individuals together one at a time.

To see this type of clustering with an example:

Let’s say we have 6 individuals - a b c d e f 

In the first step we can have b and c grouped  together and d and e grouped together.

So now we have a bc de f.

Next we join f with de. So we have a bc def. 

Next, we join bc with def. So we have a bcdef. Finally, we join a to get abcdef.

In this way, we can slash the clustering solution at any level to get clarity.

Finally, we have to be careful about the individuals belonging to the right cluster and the data integrity so that we have the right span of data ranges for the analysis.

The step 3 in creating User Personas is Building Archetypes.

Building Archetypes

Archetypes are basically the visual representation of the data we collected in the form of survey, web analytics of users or follow up open ended questions.

It is recommended to start building the archetype from user needs.

Creating user personas also help in unifying the whole organization around what is needed to satisfy the user needs as there is clear identification.

From cluster analysis, we only get the starting point to build a complete archetype in the user persona.

In the car buying example, if a person gives highest value to value of the car, followed by accident history and finally the fuel, safety and environment. This is the only information available to us.

We do not know much about the demographics. 

So, in archetype building, we go deeper and identify the demographics like gender, age range, where the person is located, marital status, salary, etc. 

We need to include demographics to help create the archetype but not too many as it is not required.

Next is we choose an image for the archetype that quite resonates some of the features of the demographics.

This helps the company in connecting and empathizing with the real user.

We can then pull a quote from the open ended questions we asked. This helps in connecting.

This also helps in validating the quantitative data we collected and also helps to empathize.

Conclusion

Now, when we are done with the three steps of collecting the data, statistical clustering, and building the archetypes, we now have to pull them all together and communicate the results.

In making the document, some of the guidelines to be considered is to make use of different visualization tools for the quantitative data.

We can make use of some diversity like graphs - different types, word cloud - where the most important word is in larger font, 

One more thing we can include is the user journey. This will explain the various activities of the user in the buying journey and the duration.

One last thing to keep in mind is that the user personas once created, can be put to use from several months upto a year.

As with time, if the factors like user preferences or even the company’s value proposition and services or customer type changes, we need to update the user personas accordingly.

But unless there is a major competitor in the market or some major factor, there won't be much changes and can be easily used for a year. 

One industry user personas can also be used as hypotheses for other industries.