Studies are usually run to validate certain hypotheses or answer an open question. However, one study’s results can often be employed to answer multiple questions. Actually, the product in question doesn’t even need to be the same. Insights that
The underlying problem
Assuming that there are multiple teams or just different people within your company who run research, you quickly lose the overview of all the data that has been gathered. While one reason for this problem is the fact that we have limited mental capacity and tend to forget things, not having a defined system for categorizing and storing data contributes to the loss of information. On top of that, problems with communicating data arise quickly when data comes from different sources. Departments within one company don’t always communicate and collaborate freely for a number of reasons.
Bad Memory: once a question concerning a product or its users arises within a company, whoever is responsible for answering this question will start to look for existing data. Even if the person looking for information assumes or knows that this topic has been explored, they have to reach out to every possible contact person first. Even if said contact person remembers which project could help to answer the question, they probably won’t be able to reference the relevant piece of data directly but instead will send the whole project report.
Impeded communication: communication via email, phone or sometimes even face-to-face is often inefficient. Especially if it’s unknown who exactly is the right contact person a lot of time needs to be invested in communication. It’s not only the person looking for information who loses time during this process. Every single person that has been contacted needs to deal with the problem, look for appropriate data and send an answer.
Nonexistent archiving System: having finished a research project, the researchers will write a report, archive it and send it to relevant stakeholders. However, these reports are hardly read thoroughly. As those reports tend to contain a great number of relevant insights–maybe even surpassing the original research question–it’s often more time intensive to go over multiple reports looking for some piece of information than to run a quick survey or user test. The result is unnecessary duplicate work.
Multiple sources of information: As we have stated often before, there are multiple departments within a company that have data concerning the target group and their behavior. Not only market researchers and the UX department but also customer service, marketing, and product management might hold information concerning your customers. On top of that, nearly every research project will have a number of secondary results that weren’t their main focus. As these are not mirrored in the title or summary of the report they can be extremely hard to rediscover.
Make data more discoverable by using its smallest unit
To make archiving your data and searching your insights more efficient you will need some kind of central, structured database. To make it searchable it’s important that you don’t save insights as full reports. Even if you give your reports appropriate tags for every piece of knowledge you discovered, you still have to face the problem that you have to search for pieces of information within the report. A more sensible approach is breaking up reports as much as possible and saving those smallest units of insight directly in your database.
This approach is called atomic research. The so-called atomic units or basic units of research are discoveries that are supported by one or more pieces of proof and tagged accordingly. Units should describe the knowledge that hab been gained or “what did we learn?”. The evidence for the gained knowledge can consist of users’ opinions, behavior, survey results etc. Tags can describe characteristics of the user or product that has been tested as well as a description of what happened or of the event was positive or negative.
The goal of Atomic R
Atomic research is related to atomic design as well as the DIKW (Data, Information, Knowledge, Wisdom)-model. The DIKW-model assumes that decisions are made based on wisdom which is derived from knowledge. Knowledge is gained from information and the smallest unit of information is data.
Atomic design uses modules which facilitate design processes. For example, when designing a website you don’t have to start from scratch every time but can compose designs from smaller units which can be flexibly combined and reused across projects and are more efficient in the long run.
To sum it up, atomic research aims to gain insights from a number of small research results that can be combined freely. If you work this way, facts from different sources can back your assumptions and multiple assumptions can lead you to a conclusion. On top of that data and insights can be connected–for example based on similar or compatible tags. Assumptions and conclusions become more valid as they are backed by a greater number of facts.
Right now the ideal solution for storing, tagging and connecting atomic research data doesn’t seem to be available. However, if it existed, a system that lets you store research results in a searchable, collaborative way would enable companies to save a lot of time and money.
We’re developing a solution for efficient and collaborative storage, tagging and evaluation