Data democratization: Tools and design practices around the good use of data

Humanity has always considered information important. At first, with the intention of keeping record, the information was transmitted from voice to voice, then, with the writing, it was registered and stored. Today information, understood as data, has acquired a new level of importance: It is relevant and an essential part of decision making in small, medium and large companies.

However, small companies are those that tend to excel for their agility in making and executing decisions. So, why is this phenomenon? Does the size of the team influence? Does it have to do with the amount of data to be analyzed? Or will it be the result of generating specific services and / or products, instead of diversifying?

Data and service design

The truth has to be told, many of the answers have to do with the agility of small businesses or startups: The communication between teams and all the information that each one collects, analyzes and generates are the factors that indicate the direction that each company must take, how to improve the products and services to offer, but to a greater number of interlocutors and data collected, less speed of action and reaction.

To equalize speed, larger companies have to take advantage of the storage and information processing capacity, making the data worth gold in the industry.

How is total coordination achieved in the company’s processes? Keeping all members aware of the information necessary to be competitive. From here lies the importance of democratizing data, beyond the desire to allow information to reach all concerned, it is about everyone having access to the information necessary to carry out their work.

We are talking about generating products and services from data collected and analyzed by each company, creating their own stability and growth. A practice that has been strengthened with the formalization of the design of products and services, as well as contributions of methodologies such as design thinking.

Figure 1: Data driven product / service design: In each phase of the service design, the data can be extremely useful to improve the proposal.

Data to create and scale a product

We can divide the creation process and product scaling in 3 steps: Generate an idea, turn it into a product and improve it or diversify its market. Certainly, each moment benefits from the relevant information collected, allowing generating, delivering and capturing value, which gives a clear clue as to how useful the data is to optimize the offer of each company, facilitating decision making in the customer relationship — brand — product.

Another significant advance that has allowed to take the data to its privileged position, is in the technologies that allow to store and process large amounts of information, besides creating and implementing an infinity of machine learning models, automating and optimizing the analysis.

Tim Brown, IDEO´s CEO and one of the most relevant voices of design thinking (apart from being the creator of the methodology) comments:

“New tools such as artificial intelligence, the internet of things and biomimicry mean our design ambitions are limited only by our imagination. Meanwhile, creativity has never been more important. The global economy is stuttering and disruptive technologies challenge established business models.”

Brown reinforces the position of the article: the importance of the data is beyond traceability, but also questions us if we have the capacity to respond with creative proposals, at the same speed at which the information arrives. Here we find one of the main reasons why a good number of companies have not yet adopted these data democratization practices: paralysis by analysis.

This phenomenon is about not having enough time to analyze large amounts of information and find insights of value, which lead to propose new products and improve existing ones. Not being able to trigger quickly, we decided not to design from the data. This happens when we maintain a linear production scheme, like the one shown in image 1.

It is then when the proposal of a circular design, like that of Tim Brown becomes relevant. In this scheme, we no longer collect information only at the end of the process, but we collect it and analyze it in each of the steps and for specific purposes. In this way, the volume of data analyzed in each step is smaller, reducing the possibility of paralysis by analysis. So we convert products into services with incremental value, as shown in image 2.

Figure 2: Data driven circular design: By focusing data collection to the specific end of each phase, we can avoid paralysis by large amounts of information and constantly offer improvements on our value promise.

The DriveNow case, a car sharing service that started 8 years ago in Germany and currently operates in more than 9 countries, with a fleet of more than 5900 vehicles, turns out to be a good example of circular design, where technology, collection and data management are well applied.

Since its launch, the carsharing quickly became a joint venture between Sixt and BMW, where the last company stopped focusing on its flagship product (vehicles) and aimed at diversifying the business, ensuring that the entire fleet of carsharing is its own and collecting data continuously on: routes, schedules, preferences and relevant information that improves the trip.

Benefits according to data type

Apart from the problem of paralysis by analysis, it is also common to find situations in which companies lack large amounts of data, especially in areas apart of historical sales.

This situation is easy to understand, if we bear in mind that the most relevant figures for a company are related to production costs and profits, but it turns out to be a priority problem when we realize that we have no answer to: which segments of high value I am leaving unattended? or who can be a representative figure to promote my products / services? These questions could be solved with data, beyond sales.

We can see that the key to successful data democratization is establishing a collection culture in all areas of the company.

Image 3: Data, like labor disciplines, generate value when they are crossed with each other.

Questions and visualization as media

The variety of data collected is as beneficial as the different looks of the areas when analyzing information. The point is that the objectives of each one can vary, making communication between them difficult. This is one of the reasons why small companies tend to be more agile than larger ones.

Hence the importance of a third conditional, to achieve a successful democratization of data: you must have an effective communication with those involved, the goal is that we all speak the same language.

Here are some tips to achieve fluid communication:

1. A periodicity should be established in the data collection, otherwise, it will be impossible to make the correct crossing of the information with its new versions.
2. Each stakeholder should clarify what they expect to find in the data, since each area benefits from the information collected by the others.
3. Share the data through visualizations. For this, we must ensure that the graph speaks for itself, summarizing and hinting at the results of the information analysis.

Visualization, key element in data democratization

Dont forget that those involved are the same spectators of the reports, for this reason, we must ask the following questions to understand what they expect to see:

1. Who will visualize the content?
2. What do you want them to know or do with that information?

Image 4: For a correct data visualization in an explanatory analysis, which seeks to tell a specific story, it is advisable to start by answering the 3 questions: what? who? and how?

The answers to these questions help us, as visualizers, to select the appropriate displays according to what we want to communicate.

• Simple text: It is the best way to communicate when there are only one or two numbers to share.
• Tables: The data should always stand out, so the tables are ideal for communicating different units of measurement.
• Heat maps: A mix between table and visuals. We can facilitate reading with the help of nuances and saturation of colors, from a general view to the detail of the information.
• Scatter charts: They are useful to show the relation between two subjects, since the location of the points with respect to the axes shows the relation.
• Standard: The standard graphics are the most used to show continuous data.
• Slope Graphs: Es la mejor opción cuando buscamos comparar aumentos y disminuciones relativas, entre dos puntos de vista o dos periodos de tiempo.
• Bars: Vertical, horizontal and stacked, are usually a type of graphics quite recurrent when it comes to displaying information.

Now, if each graph has the potential to transmit a specific message about a piece of information, a good report has to work by telling a complete story, for which is convenient to follow these steps:

1. Sketch: We tend to underestimate the potential of paper, but a good session of sketching, in which the final users of the visualization put on paper their expectations, helps to understand what data is looking for and how they want to visualize them.
2. Mockup: Once we have a sketch on paper, we can make the data model to visualize the selected tool. At this point we focus on functionality.
3. Final visualization: At this point we focus on the look & feel: colors, typography and other visual elements.The concern should not include essential information raised in the sketch, offering a nice and clear visualization.

In case of being a dynamic report of periodic delivery, we can propose a roadmap of improvements.

In short, the data democratization in an organization becomes an essential effort to streamline decision making, basing them on a growth strategy that offers value to its customers. With this, it is possible to maintain a clear and continuous communication with the final users of the products. We finish with a step by step process:

1. Define the data that is collected in each area, in order to clarify the purpose of the collection of each source.
2. Establish an internal culture of data collection and review the possibilities resulting from crossing information.
3. Define the visualizations that each one of the involved parties can offer, thinking about the periodicity of said reports and what the viewers will look for in the data.