Creating Data Products People Use
How do you ensure that business users get the most out of analytics? How can you drive a data-driven culture? The most important step is creating a solution that is easy to use. In this context, a data product is a tool or service that facilitates an end user’s interaction with data. This could take the form of a dashboard, report, automated flag, presentation etc.
Understand End User Needs
Data products typically exist to answer questions and monitor metrics. They are there to find answers to business problems and as such, your end users are key stakeholders throughout planning and implementing. Data teams need to leverage the knowledge of their subject matter experts in order to understand the questions they are answering, the decisions they are affecting and today’s process against a more data-driven one.
We find it is a good idea to engage people during your ideation phase and find out what a data product might look like in a ‘perfect world.’ and try and work backwards.
Contextualise End User Needs
Once you have an idea of what the product might look like in an ideal world, you then have to consider today’s picture and understand the gap between those two. It sounds obvious but, your data product needs to work. It needs to work with a high degree of accuracy, timeliness and consistency.
Sometimes you will need to work around the limitations of your current tools and available data. Sometimes you will need to work around the current skill set of your analytics team. In some cases it might be better NOT to work around things and only deliver what you can.
One example of this would be Linkedin. Former head of data products DJ Patil said they had considered a feature that would market jobs to people with matching skills, but realised that approach was fraught with peril.
“There was too high of a chance that we’d accidentally recommend a more senior person should apply for an internship. When stuff like that happens, people get pretty pissed off, and it can mess up your brand pretty fast.” While this feature exists today, at the time, it could not be executed properly and so it was better to hold off. Consider how you can accomplish your ultimate goal in the simplest way possible.
Plan Iteratively
Start with an MVP (minimum viable product) and then gather user feedback, track usage, track ROI or outcomes. From there you can plan out a product roadmap that will eventually take you to that ideal version your stakeholders had in mind. This approach allows for flexibility and responsiveness to user needs and can help you to plan for scalability issues such as increasing data volumes and user loads. It also protects your team from working on an initiative that people are not ultimately using and can allow you to pivot to say, a different dashboard tool if end-users are struggling with the current one.
Make It User Friendly
Ensuring a user-friendly interface is paramount for boosting product adoption rates. By prioritising intuitiveness, ease of navigation, and visual appeal, you pave the way for a seamless user experience. Simplifying complex processes and minimising the learning curve can significantly enhance user engagement and satisfaction. Introducing new tools can present hurdles, requiring additional training for users to effectively utilise them.
As such, integrating your data product into existing & familiar tools can mitigate resistance and streamline adoption. For example using reverse ETL tools to create flags or workflows inside of a CRM tool. The goal is to minimise friction. Integration with familiar tools not only reduces the need for extensive training but also enhances productivity by allowing users to leverage their existing skill sets.
Quality Assurance
The most important part of any data product is ensuring you have clean and validated data. Poor data quality can lead to flawed analyses, misinformed strategies, and ultimately, costly mistakes. Data quality is also essential for maintaining trust among stakeholders, whether they are internal decision-makers, customers, or partners. Additionally, data quality may be a prerequisite for compliance with regulatory requirements.
Training And Support
Never underestimate the amount of training and ongoing support that will be needed to ensure that end users are using your data product. Try setting them different tasks and challenges to ensure they know how to use different features. Work in a ‘classroom environment’ where you can see how people use the tool and guide them as opposed to purely online training or videos.
Plan in extra sessions or ‘top-up- classes where necessary and try to identify which team members get to grips with the tool easily, as they might be prime candidates to become Data Champions and help others to adopt the data product.
Feedback
Establish a feedback mechanism to collect user input and continuously improve the data product based on user suggestions and requirements. This could be as simple as a link to a survey inside a dashboard.
Conclusion
All data projects fail if no one uses them and so you should spend as much time worrying about adoption as you do in creating the solution.
The challenge for data teams will always be understanding user’s needs and guiding them to what solutions will work today, planning for tomorrow, and ensuring ongoing adoption.
If you are looking to create data products that your team will love and will drive real business growth, why not get in touch with the friendly 173tech team today?