Performing a market research using your app user data is always tough, but incredibly useful, specially because technology and methodologies have evolved and now you can do a research on a few clicks, and you can add this to your daily tasks list with no hesitation.
This article was written by Sonia García, an stagier that was working together with our Success Dept. during 2020 summertime.
But first, you need to refine and define what you will be researching. Data analytics needs some previous steps, as you need to be aware of what you are looking for, the objectives and the information you will be getting from your users. On this step-by-step uide, we will show you how to research efficiently and how we did our own internal research using all your data.
Research workflow
Yes. Start from the very beginning, that is: planification. You need to know what your problem is or the data you need to extract, as this will help you not only when analyzing the data but to its interpretation. What am I looking for? will always be your first question.
A market research is just as any kind of research, so you need to ask the right questions to get to the right answers, specially the right answers for your business objectives. That is, if you want to know more about your competitors, ask yourself questions that may take you to a comparison: your users will probably have tried all potential solutions, so you will be getting valuable information directly from one of the best sources, specially if you are wondering about potential UX/UI issues.
Most likely, those questions can be even the same ones. And now that everything is clear enough to start, let’s move into real work. The real step-by-step guide.
Data preparation
Preparing your data analysis will the toughest and most time consuming part of your research. Period. You will need not only a person or a group of people to dig and prepare the data, but also the right questions and potential answers you want to get and some (really advanced) knowledge of tools as Excel or Google Sheets.
First things first: gather and download all the data you need from the original source, if possible. This will ensure data integrity. However, sometimes you won’t be able to access the original source or data is stored on different formats (website, survey, database), so some transformation and formatting will be needed. And, of course, filter your data. For our market research, we used user comments on our AppSumo deal.
Good filtering is not only one of the issues you will be facing, as there is another important challenge: dataset can not only be huge, but also have a large number of variables, so remove or archive the ones you will not need for your research.
Once you are ready to move on, we have a recommendation for you: copy and paste all that filtered data onto a new sheet or, better, a new document. Doing this will help you working, as your will have not only the data you need but als owill free some resources on your computer, which is critical if you are working with a huge database.
And if you feel that you are overwhelmed, you may review your filters or, better, split the data into variable groups, so you can get better and clearer information gathering all related data. And, of course, keep the order: if you are tagging your users using e.g. their company name, use that variable to keep everything in order, as this could alter your final conclusions.
To get the better results, you need to understand any piece of data and what it refers to
Finally, if you are processing your data using any analytics or statistical tool such as SPSS, encode your data. What is this?
At this point, if you have been working with data for a while, you may have learned a lesson or two about data, letters and numbers: caps, points, apostrophes can make your data a mess. If you convert all the gathered information into numbers, your data processing will be less painful. Let’s see it on a brief example.
If you use this simple conversion, male=0 and female=1, SPSS will be able to get more information and, therefore, you will be able to get better conclusions. And you will be able to combine that information with some other variables.
But, what’s best? Using letters or numbers? We can’t tell you. Convert and transform the information to the format that is more comfortable to you. Some data would be better to understand using an scale, otherwise you’ll find yourself reviewing nonsense information.
Encoding may seem simple, but it’s not that simple when working with user data. Users make mistakes on inputs, and typos are everywhere. So, take your time to review most common typos using search and replace to keep your house clean.
Data processing
Data processing is simple if you know what you are working on: variables, their information, how they can help to reach to a conclusion and their type
If you know how to use your data processing tool, this task will be probably the easiest and fastest of all, as you only need to tell your computer to do it for you. You’ll be there only to review that your computer is mixing everything the way you told it so and to make your mind.
SPSS is just one tool to process huge datasets. But if you are not working with big data, you may use some other tools as your most loved spreadsheet software (Excel, Google Sheets, even Apple Numbers) to reach to any conclusion simply gathering together a few data.
And, of course, remember: qualitative information is always important. You need means, medians and quartiles / percentiles, but it’s not enough. You can’t forget in your analysis user rates, as users per continent, something that is really important here at Quoters HQ as it affects to workloads on marketing (newsletters), support (SLA’s) and even to developers (deploys).
Analysis and conclusions
Last step. If you are going to show all your conclusions to the team, don’t forget to do a last minute review. This will help you to check for potential errors, strange data and, of course, to get to know the final results of your research and explain all this in the best possible way.
And that’s all: review carefully the data, listen to them, learn what they say and reach to the best conclusions. That analysis you’ve performed can be used to make decisions.
Finalmente, si vas a presentar los datos al resto del equipo, es más que recomendable hacer una última revisión.
Esta revisión te permitirá ver si hay errores, identificar potenciales datos que no cuadren y, claro está, para conocer de antemano los resultados y poder explicarlos de forma clara y confiada. Ten en cuenta que los resultados, con amigos, son mucho mejor que decían en Barrio Sésamo.