This content, written by Colin Zima, was initially posted in Looker Blog on Aug 1, 2017. The content is subject to limited support.
Looker cuts a new release every four weeks. In October of last year, we cut Looker 4.0. Next week, we’ll cut Looker 4.20. We have always worked this way, and don’t foresee changing it anytime soon.
Our four-week release cycle is core to our engineering philosophy for two reasons.
First, we believe in giving customers the features they want as soon as they’re ready. If there is a feature, be it small or large, that a customer thinks would help them get more value out of Looker, we want to get it into their hands as soon as possible.
The second reason is data. Iterating on a feature is only possible if you’re able to work from real usage data. The sooner we can get features into the hands of users, the sooner we can know if it accomplished what we hoped it would.
The best part about moving so fast is that we’re able to use the data produced in the first few weeks of a features life cycle to adjust it and improve it so our customers get an even better version of a feature four weeks later.
The four features we announced today are just a sampling of what we’ve been working on...
The first feature we’re announcing is a real game changer in my eyes. Keeping a BI instance performant and fresh has historically been very challenging. Either queries run live, processed anew for each and every user at query time, or results are stored in some intermediate layer one step removed from the database (be that a cube or a specific result set). Administrators have to choose between fast results that may be a day behind or real-time results that may take a couple minutes to run in some cases.
Smart Caching solves that problem. By dynamically checking the cache against the raw data, the cache size can be maximized when data is unchanged, improving performance for end users. But, at the same time, when underlying data changes (this can even be set with thresholds of change!), the cache can be cleared for real-time data. There's no more balance between cache size and latency, you get both if you want them.
At the beginning of the year, we announced a set of features we call . This feature allows admins to assign attributes — anything from row-level permissions to geographic areas —to users and groups. When you add Advanced Scheduling to the picture, it means managers can maintain a single dashboard that is filtered on one of those user attributes like region, and schedule it to be sent to every user with the data filtered for that user.
Because of Looker’s flexibility and the nature of operating in-database, we’ve always handled time zones fairly well. With the new Advanced Time Zone Control features, users are able to leverage both core benefits. From the flexibility standpoint, folks can explore data or schedule reports in any time zone or even using a mix of time zones on dashboards. Additionally these can be parameterized per user or locked, so folks in different time zones can easily work on their time without thinking. Second, because we are operating in-database, there's instantaneous ability to query in any time zone seamlessly. No rebuilding data pipelines or anything, just simple at-query-time conversion.
Visualizations are key for telling a story with data. The more visualization types we can give our customers, the better stories they can tell with their data. Because of this, we added a funnel visualization and a timeline visualization.
More features are always rolling, so keep an eye out for more updates. Our user conference, , is right around the corner, and we all know conferences = announcements!