There are millions of data points ready to brighten your day. The glass is no longer half empty; it has broken. Just as information overload can give us the temporary satisfaction of enlightenment, accumulating data in bulk adds to the list of hobbies that satisfy specific patterns where quantity prevails over quality.
Globant’s Create Studio orchestrates all marketing areas fueled by the intensive use of AI and data. When you design a dashboard, everything is about asking a specific question at each step.
What is my business goal?
The right questions for dashboard creation should always come first; otherwise, it builds on nothing. What is my business goal? How am I going to accomplish it? How am I going to measure it?
The result will always be dashboard data. My business objective will be to sell more books with golden flowers. I will achieve this goal by starting with a creative communication campaign and measuring my final profit.
What is my plan to measure each phase?
Metaphysics does not convert on Facebook. Sunglasses at 10 euros, yes.
When tackling a job with data, a roadmap is crucial, and it must start with the goal. Only go to the data after you know what you want to find, or you will end up working to gather data and not looking at everything from above, knowing what you need to look for for your plan.
With the business objective, it is time to go down the different phases of the communication plan, and at each phase, a measurable and actionable objective (KPI) that contextualizes the upcoming actions.
If we want to sell more books with golden flowers thanks to Instagram, we must be clear about the budget, the costs per acquisition, how many impacts we need to convert, and how effective our creativity is.
What should my KPIs be?
The answer is simple: that data is actionable for the business. The rest can generate more noise than help.
See how the business can change with the data, not vice versa.
Avoid pivoting on vanity KPIs, fans, page views, and impressions. Actionable KPIs provide quality data.
How do I know if something is good?
The perception of judging everything from a personal perspective, without data analysis to anchor it, has more risk than success. Reduce the uncertainty of your decisions based on the search for internal and external patterns.
Before launching the campaign, start with reviewing trend reports, other campaigns, studies, benchmarks, and case studies carried out by important platforms, specialized companies, and advertisers.
In this way, personal taste will have data to anchor it. Based on the successes and errors of third parties, you will avoid a lot of trial and error.
How do I configure the extraction of my data?
Once the campaign is underway comes the point of data extraction. Having data will fully enrich all decisions. This step is vital because it will mark the effectiveness of the following ones.
If you misconfigure the extract, it will lead to chaos when you go to model the data. And it will be time to start all over again from the beginning.
A key point, take good care of how you keep the raw data. It is your most significant value. Both to keep it in the future and be independent of the source or connectors and to be able to check data 10 months later in case of any doubt, without anyone having deleted or manipulated the original.
If a doubt arises about a piece of information, you can start to doubt everything and throw away months of work.
How do I give context and utility to the data?
With context, the data is clear. It’s good that you have 100 million impressions. But is it so good if you are spending 1 million euros to get them? Maybe not so much. Or yes, depending on the context.
Each metric has its partner or large family to contextualize data. For example, the global figure of 20 million views in video views is very high. If those views were from paid and only saw 1 second of the video, the quality of the views is nil. In this case, the percentage of video viewed, the VTR by sections, and the seconds are fundamental.
What is the importance of each data point?
Not everyone needs to have the same level of detail about the data. Establish different reading levels according to the profiles of those who will read the report. Each will want a piece of data to execute for their business area.
From there arises the pyramid of importance. Through your business objective, achieve a categorization that helps you filter the data. A funnel helps organize ideas.
How can I visualize thinking of two data points and not 500?
Once extracted, ordered, and filtered, you need to be able to visualize the data. A good visualization is key to finishing off all the efforts made previously. A bad visualization will destroy the best work.
Again, think about the least amount of data, how you could activate it, and how, at a glance, someone who needs to learn what that data is about would be comfortable with it.
Context is critical to good visualization. Help yourself with different graphs and tables that make a whole story. If there are any graphs or points along the way that make the story confusing, delete them.
Will my data help to make decisions?
A report that does not help to make a decision is not useful. Spend as much time as possible extracting all possible insights to make those decisions actionable and worth all the effort spent. These insights will improve the roadmap, helping you optimize it and benefit the business long-term.
When does my data strategy end?
Short answer: never.
Now it is time to repeat the process over and over again. Start as many times as necessary because if you have followed the entire process, on the next occasion, you will no longer start from point 3 but from point 40, thanks to your data and conclusions.
The uncertainty will be less, although it will never disappear, and with each new step, you will improve the whole process until it turns on its own.
Learn more about Globant Create and how to reach the next level of growth in your marketing efforts.n your marketing efforts.