Big Data for the Small Stuff
By Rachel Obstler | March 30, 2015
I love data. I’ve had many jobs throughout my career where I was analyzing, presenting, and selling data. Similarly, at Keynote, the data we provide to our customers so that they can improve their customers’ digital experience is at the core of our value proposition.
Big data became the new buzzword back in 2012, and to an industry like ours, it seemed like serendipity. Since then many companies have launched analytics products, many of which allow unprecedented access to large volumes of data and the ability to build customized dashboards and query whatever you want, whenever you want.
However, like any of the technology trends in recent years, there is a gap between promised vision and reality. The promise of big data is not about putting a business intelligence tool in front of a database and telling customers to have at it. Business Intelligence (BI) tools certainly make it easier to get to an answer quickly across larger and more varied datasets, but knowing what to ask is often a much more difficult task, which BI tools don't help address.
Keynote has just launched the second product on top of our new big data architecture, called Digital Performance Intelligence (the first being RUM—see “Get the Inside Scoop: DUM vs. RUM”). This release of Digital Performance Intelligence answers the specific questions “How does my website compare to other websites” and “How are my third-party services performing on my site vs. others, and vs. other third-party services in the same category?”
You can infer a lot about how other companies have architected their best of breed sites through knowing what gets loaded when or which third-party services are being used. This product is essentially a visual representation of web content, design, and performance for the performance analyst. It’s not designed to diagnose a one-time incident. It’s intended to surface systemic performance issues over time that may be difficult to discern because they are caused by one small component, or high variability, or are otherwise not easy to spot by looking at a single waterfall chart.
We’ve also leveraged big data to provide third-party benchmarking. The use of third-party services is ever increasing due to the need to integrate social networking, analytics tools, advertising, cloud computing, or other non-core services into websites. Keynote has millions of measurements running each day, and we aggregate each third party’s performance across all of these measurements. If you find a third party causing customer-impacting performance problems, you’ll not only be able to understand how they are impacting the customer’s experience, you’ll also be able to find a better-performing alternative.
Big data and all of its potential is awesome. It’s allowing us to quickly average and trend performance, down to the smallest component level, across thousands of websites, in real time. But it’s not the promise of scale and speed in and of itself that makes big data so interesting; its promise is fulfilled only when we successfully find a specific customer need that it can serve.