One of our greatest concerns when starting BTBuckets was scaling our infrastructure. We needed both cloud and platform services to fit our needs, and we tested many cloud service providers. In the end, we decided to create a mixed environment, running the profiling engine on App Engine and the front end on Amazon EC2.Our experience with App Engine was smooth. There were times when the datastore would time out too frequently, but it was manageable and the application kept them from affecting our users. The biggest gain with App Engine is not needing to maintain our backend servers anymore. The old model, whether with in-house servers or an outside data center, involved disaster recovery and other high costs of maintaining infrastructure. We are relieved now that we can just focus on our business.
After creating a profiling and targeting framework with a small footprint and a very low cost per request, the next step was to integrate it with two great Google products to improve sales and conversions: Google Analytics and Google Website Optimizer.
Google Analytics helps us understand website navigation and usage, while Website Optimizer allows optimization via A/B testing. Built on top of them, BTBuckets and its Chrome plug-in allows you to create user segments right within Google Analytics, as well as to apply A/B testing to a subset of the website audience. The user might even go beyond that, by easily creating custom actions for each created segment.
We believe BTBuckets is a disruptive product: it adds value to the most common web analytics tools, it's generic enough to allow any customization through the JavaScript API, and it's free to use.
If you're attending Google I/O, be sure to visit us at the App Engine pod in the Developer Sandbox to learn more about BTBuckets!
By Arnaldo M Pereira, BTBuckets team
No comments:
Post a Comment