Data analysis: do you need a dashboard?

To make this data speak, figures relating to performance indicators, such as the return on investment, the cost per acquisition, or the margin generated, are extracted from it. But to have the complete history of the data, it is necessary to link these data together: we can thus know that the results of the X campaign in Facebook correspond to X sessions in Google Analytics, and generated X dollars of sales via the internal information system. To achieve this, it is necessary to unify the data, that is to say, to know that the identifier John in the database 1 corresponds to the driver's license number LFURKF67520.

Since the 2010s, the emergence of data visualization and unification platforms (thanks to connectors making links with other platforms) such as Datorama, Tableau or Power BI have greatly democratized the practice of dashboards, up to then reserved for insiders of the code only. Admittedly, their use is now open to a much larger pool of users than before. However, it would be wrong to claim that anyone can correctly use these tools without having any notion of data modeling. Making effective dashboards requires thought and planning ahead of their creation. Here is an overview of the major issues in the practice of dashboards.

 

Mainly to save time! Doing  campaign reporting  is tedious because a lot of manual tasks are involved. In addition, the collection must be restarted with each report because the data is frozen in time. Thus, creating a report once in a visualization platform makes it possible to automate this task, and therefore to spend more time analyzing the results.

But beware, these benefits come with some pitfalls. The main risk would be to waste your time on technical “plumbing” rather than analyzing results. Like an automobile, the costs associated with a dashboard are not limited to its creation. Maintenance costs are to be expected because the raw material, the data, evolve, and therefore periodic adjustments are necessary. Adjusting campaign names, reviewing the formulas that manage data groupings (channel, tactic, segment, etc.) are all tasks that can encroach on analysis time.

The data structure and processes in place are of paramount importance to the success of a dashboard. On the one hand, they allow operational efficiency that makes the  dashboard  credible: displaying accurate data. Without maintenance, a dashboard simply becomes obsolete.

On the other hand, costs related to maintenance can be very high if no governance framework is defined.

 

The following two points are important in order to maintain the effectiveness of your  reporting.

 

AUTOMATING REPORTING IS NOT ENOUGH TO AUTOMATE DATA ANALYSIS  

A big misunderstanding on the subject of data analysis and its representation lies in the fact that many confuse two tasks,  reporting  and  analytics , which do not have the same objective at all.

1. A reporting platform   is used to automate a company's  reporting  , that is to say, to make data easily accessible, in the form of coherent sets.
2. Analytics  is about exploring data and extracting insights  from it .

The majority of organizations want dashboards to extract  insights from . It is therefore analytics  that they need and not  reporting.  Your KPIs should measure the health of your business goals, not smear your data. A dashboard that is too detailed and without KPIs, with only metrics, will be more like  reporting  than analytics , and will then be useless to extract  insights from  it because they will be hidden behind the tide of figures put from the front.

 

THESE PLATFORMS SHOULD SAVE YOU TIME, NOT THE OTHER WAY AROUND: THE VARIABLES AND PROCESSES TO CONSIDER

Are your KPIs chosen based on your business objectives? Now is the time to segment them. But beware! How far do you need to go? It should always be kept in mind that creating a dashboard has two objectives:

• Bring up the right questions . A dashboard rarely gives answers, a good dashboard allows you to raise relevant questions that will help you better understand your performance.
• Save you time . The danger here is to want to do something too complex, which in the end will be too granular to bring real added value and possibly drain all your maintenance time, to the detriment of analysis.

 

WHAT ARE THE BEST PRACTICES TO FOLLOW?

It is important to think about the basis of performance measurement: what do you really need to evaluate your performance? We are talking here about KPIs and not metrics, which you should choose 2 or 3 in number. Then, segment these metrics by a few main segments, and the architecture of your dashboard is done. This exercise is usually done during your KPI framework.

Once your KPIs have been chosen, you must select the metrics that will break them down and explain them. Here are some ways to choose them well:

• Always keep a link with business objective;
• Keep it simple;
• Choose metrics that drive action;
• Be transparent: a report whose methodology is specified is never false.

Finally, you have to choose the right type of dashboard according to the audience it is aimed at, because the dashboard that answers all the questions of everyone does not exist. Your dashboard will be very different depending on whether it is used by your management ( strategic dashboard  ), whether it is used by your managers to evaluate your campaigns ( tactical dashboard  ) or to monitor your operations ( operational dashboard  ).

The democratization of the tools used to create these dashboards now makes the practice accessible to everyone. By following the above advice, any organization, from small business to large group, can effectively assess its performance against set objectives, and thus take action to improve.

Good analysis!