5 Weapons to Give your Data Visualizations Maximum Impact
The recent rise of digital has influenced many businesses. One clear result is the massive amount of data generated. Incredibly, it should reach 163 ZB by 2025 according to the recent report “the Data Age”. That is ten times more than in 2017. Companies have recently realized its potential as a source of growth and have started adapting to take advantage of it.
We can now see new kinds of personalized marketing shaping relationships between brands and consumers including the arrival of products and services which are in a constant evolution based on user signals like Netflix or even more indicative, digital-native companies building their business model on the resale of their data.
However, to profit from this new income stream, organizations first need technical and analytics expertise to reveal the insights in a large amount of available data. Very ofter, the storage format is a double-entry table, each cell provides a piece of information which is related to a variable (column) and an observation (line). When it is composed of more than a million lines, it cannot be read and understand as it is. To benefit from its insights, data must be cleverly visualized and not suffer the common mistakes and bias interfering with its comprehension, but how?
1.Restrict the amount of KPIs
The first key is to restrict the amount of KPIs present in your dashboards, ensuring a focused approach. Very often people tend to fall into the ‘data rich, information poor’ syndrome and try to use the most possible the data into visualizations. This is called a report, not a dashboard. It is a summary of large tables providing no help in the decision-making process.
Think for a moment about Google analytics reports. Do all the indicators help the evaluation of a website’s performance? Unfortunately not. Assessing a website performance using the all report equal looking for a needle in a haystack.
So to succeed, you have to be selective. Only show indicators related to business objectives or the key factors influencing them. It involves subjectivity but it is also your added value to know which information is needed and which is not. A dashboard gives a clear big picture of the company’s performance.
2.Adapt the information granularity to the audience
The dashboard information granularity needs to be adapted for the audience using it. Consider a media trader, running ads for an e-retailer for example. The trader’s top executives (CEO or CMO) will only be interested in the ROI coming out of its online campaigns whereas the traffic acquisition manager will spend more time investigating which campaign has performed better, on which audience etc. So before trying to add every kind of data available, think about how helpful it will be for the user. Based on its level of knowledge does it have the ability to treat and use it or, quite the opposite, will it cause it to drown in a deep ocean of information.
3. Find the right tool to visualize your data
Still upstream from the dashboard conception, a list of data resources used for visualizations must be established. Some might require engineering operations, others not. The real issue is to foresee the questions to be sure the dashboard is working well when it is needed. According to the tool used for visualization, some native connectors can exist. These can allow you to used pre-aggregated data from most common APIs like Google Analytics, Business Manager, Google Ads, and many others without processing the data into tables.
For example within Google Data Studio, there are plenty of native connectors. It is possible to use datasets from Bigquery, Google Analytics or Google Sheets and instantly have all the dimensions and metrics available, making it very easy to manipulate thereafter.
4. Follow a coherent structure
Once you have an exhaustive overview of your different KPIs and data sources, start designing a structure that will help the user comprehension. Dashboarding is like telling a story, the reader has to understand the logical consequence of each step he has to go through. A key to give visualizations impact is to drive its reading through the different steps that are consequences to one another. Is your company performing well? What are the key factors that lead to this level of performance? What are the insights and the context you can link to it?
A very common example compares a business dashboard with a car dashboard to understand how to structure the different parts. The speed is the KPI of success, in this case, 100km/h. Face it with your objective which is to reach 130 km/h. The first message to highlight is the performance is not satisfying. The second step is to show the key metrics leading to this result. In a car, it would be represented by the different alert symbols. Only one is lighted in our example, the engine one. The last steps consist of presenting the relevant report section explaining the different figures gathered before and spot the source of the problem. This would be translated into a mechanic report about the car engine.
5. Provide context to the user
Once the content of each part is clearly described, the dashboard needs to be brought to life by adding context and emotions to it. Indeed numbers, even the most astonishing ones, without context, remain numbers. To generate discussion and decision making, it is important to add comments, feedback or anything that makes your numbers relevant and meaningful. Recalling the objectives or one particular event that has influenced your performance will be always appreciated and help comprehension.
You notice a strong click-through rate increase in your report and realize it is directly related to the new set of creatives ads you have recently run? Therefore you could show the high performing creative content in the dashboard. It might be the fertile ground for debate and future development.
You spot the number of sessions drop compared to the previous month? You know it is directly related to your website 3 days shutdown, then you should add it as a comment below the graph.
6. Free yourself from bias
Finally, pay close attention to any kind of bias that can interfere with user appreciation. Make sure you don’t, willingly or not, underestimate or overestimate your data, even if the performance is collapsing. Sidestep the message using the scale, ratios, or any irrelevant information to give a brighter picture is not part of your mission, especially as it aims to help decision making and lead to strategic decisions.
Revealing insights from a database requires tools, methods, and stance. The way data is visualized depends on the analyst. By clever use of the available tools, the reader’s comprehension of the data can be hugely improved and you can deliver clear, usable and targeted messages with impact.
Do you have a favourite tool or technique? Share it below and, together, we can make those impacts even better!
About this article
This article has been written by a student on the Grenoble Ecole de Management’s Advanced Masters in Digital Business Strategy. As part of a content creation assignment, students are given the task of writing articles based on their digital interests and disseminate the articles online. Articles are marked but we make minimal changes to the content. Thanks for reading! James Barisic, Programme Director, MS DBS.
