Data-aided decision making is on the rise. In 2015, IDC reported 22% of enterprises created advanced digital transformation initiatives to better respond to customers. However, that number is expected to double by 2020 as more companies begin to see value in data-driven initiatives. Provost and Fawcett explored this approach ("Data Science and its Relation to Big Data and Data-Driven Decision Making"). Citing multiple real-world examples, they conclude data-driven decision making is in fact beneficial over intuition alone.
Leveraging data to support or counter a decision can be challenging, even in the age of big data. Website development is often unequal parts of coding and managing expectations through a series of meetings. Any one meeting with stakeholders can produce a feature request for elements that may not be a good fit for the site. Gut instinct enters the picture when both developers and stakeholders are certain a feature is perfect (or not) for the site. Delving into existing site data may be the only way to choose the proverbial next step.
Admiration of a competitor's web presence isn't the only motivation for requesting new functionality. Organizational politics can also drive change requests that could prove to be time consuming, costly or both. Taking time to fully understand the request can, of course, enable developers to consider the needs of internal stakeholders along with the needs of end users. Asking the early questions about long-term goals and perceived benefits can help define a potential use case for the feature.
Moving quickly to add timely content or trendy functions can be beneficial. However, existing visitor data may reveal a much more urgent need for improvement elsewhere. Analytics data, particularly Google Analytics, can show the path visitors take from one page to another. For example, a sudden drop-off in page views or pages that consistently lead to support inquiries may indicate friction between your call-to-action and contextual content for the action. On-page analytics may show a usage pattern that is contrary to a particular call-to-action or interactive element.
Asking the right questions of your data is crucial. Start small; don't analyze everything at once. Determining which metrics are relevant, and formulating questions around them, will help define your plan for analysis. Using website content as a test example, you may want answers to the following:
- Multiple product pages and call-to-actions are being A/B tested. What generates the most traffic to the help desk?
- A video for new feature y generates an increase in sales questions, but few sales. What additional content would be required to eliminate these sales questions and shorten the buying cycle?
- Visitors are switching between mobile and desktop versions of the website. What action are they attempting just before requesting the desktop version of a site from a mobile browser?
Answering your questions with data can help your team better understand past performance and predict future actions. Your analysis can also surface weakness in current sales funnels or other technical issues not previously addressed. The newly requested feature may prove to be beneficial for a future release or unnecessary after more urgent issues are resolved. In the age of digital transformation, it's increasingly important to decide with data.