In the past decade there has been a movement for organizations to adopt a “data-driven” culture. It started as the result of technological advancements in cloud computing and analytics, making data cheaper and more accessible. Eventually it grew into an area of substantial investment by organizations with specialized data roles, various analytical tools and data that could put them one step ahead of their competition. While a theoretically sound action to take for any organization, as most soon found out, it doesn’t guarantee success. With an abundance of data at their disposal and data knowledge siloed to specialized roles, many have been left with the question of, what to do next?
The answer to this question is to adopt the next evolution of being “data-driven”, which I like to call the “annotation-driven” culture. Annotation-driven culture focuses on centralizing the flow of data insights throughout an organization by creating and storing data annotations in an easily retrievable and interpretable manner. It ensures questions about data to be answered immediately, rather than serve as the starting point for further analysis. Individuals create and store data annotations for anything from the results of a test to campaign performance and whenever a question occurs, rather than jumping immediately into conducting data analysis, the system can be used to see if anyone from the organization has already shared a data annotation that will answer the question without additional resources. Adopting this culture is not something that requires substantial investment in tools or resources, but it does require a shift in how and where data insights are communicated and stored. Most data is consumed, analyzed and shared soon after it is acquired, but in the long run most of that output is lost with time.
Think back to two months ago, can you locate or remember all of the insights you shared? How about the amount of times in a meeting where a question was brought up about a historical trend, but no one was able to answer it? You will most likely be able to answer the second question, but struggle with the first one and you aren’t alone. This happens because our memory recall isn’t large enough to remember everything we learn. It is why we have to rely on notetaking and the information from past notes to retrieve that memory. Memory recall is an issue with every aspect of life, but one area that faces more challenges than others is data analysis. Data analysis is complex due to the infinite number of data points and relationships that can be examined. Without the help of data annotations, the data analysis version of notes, any insights that are generated by data are sure to be forgotten as you have seen with the answers to the questions above.
Data is only as powerful as the insights that provide context to the data being observed. This is why summarizing your data analysis into bite sized, interpretable data annotations and storing them in one area for all members of the organization to see will help with memory recall as notes do in our everyday life. Using the steps below, I will help you to implement this culture in your organization and start to fill in the knowledge gaps presented by your data.
Steps to Adopting an Annotation-driven Culture
Step 1: Gather Feedback and Create An Annotation Process
Starting an “annotation-driven” culture at your organization begins with understanding the challenges each department faces in interpreting and sharing insights. Questions and surveys should be sent out to collect this feedback and results from the feedback should be compiled to create your annotation process. The topics you should specifically ask for with this feedback are, the common questions peers are asked about data, the description of insight sharing between departments and the platforms used to communicate insights. The answers to these topics will shape the annotation process that will be used to determine the types of insights that should be annotated and how it will be shared with others to ensure that each data annotation provides maximum value to your organization’s reporting and data analysis.
While this annotation process will be shaped by the process above, there are additional criteria that should be included, which make it helpful in navigating a database of annotations. Dates and times should be utilized so that a person observing a pattern in a given report can reference this time period within the annotation tool and quickly retrieve and understand the reasons behind the pattern. In addition, at least one system of tagging should be used so that team members not only have the ability to filter annotations by historical dates, but can also segment the annotations by the tags that are most relevant to the data they are analyzing. This tagging system will be specific to your organization’s needs and can be related to anything from departments impacted by the trend to the type of marketing reflected in the data observed.
Step 2: Find Tools to Create and Store AnnotationsWith a process in place, it is time to find a tool or tools that can be used to create and store the data annotations that are created by members of your organization. This database will act as the central repository for insights both past and present. While the tool is an important aspect of annotation-driven culture, this part should not be very time-consuming as there are an abundance of tools on the market, like CMS and project management software, to handle both the creation and storage of annotations. The only important criteria to consider when making a selection is that it should have a feed to make it easy to read and retrieve annotations, the ability to attach custom attributes to a record and a filter function that applies to the custom attributes as well as the date and time values. If for some reason you are struggling to find a tool that can do this, then start by using the extremely flexible and free combination of a Google Sheet and Google Form. These products are seamlessly integrated into each other and while not the prettiest solutions on the market, they are by far the most flexible record creation and storage systems. These tools will be the easiest to use as a sandbox version for this process until it has been fully iterated on and has grown beyond each tool’s functionality.