Improving the speed and helpfulness of our customer support operations has been a major priority over the past year. We’ve cut down the average time-to-response by about 75 percent, but we know we can still do better. To help with that, we’re moving our support software to Zendesk, so you might see some minor changes in the coming weeks.
For previous site improvements, check out all of MuckRock’s release notes, and if you’d like updates emailed to you — along with ways to help contribute to the site’s development yourself — subscribe to our developer newsletter here.
Upgrading MuckRock’s support tools
Over the years, we’ve grown from dozens to hundreds to now tens of thousands of registered users across our platforms, reaching about 80 million people each month. We’ve made a number of improvements along the way, from hiring a dedicated support team to moving beyond a simple inbox for support tickets to using machine learning to automatically handle routine request classification work.
Now we’re making some more upgrades, switching to Zendesk as our customer support tool. We’ve already made the move over for our lower-volume properties, including DocumentCloud, oTranscribe, and FOIA Machine and plan to migrate MuckRock in about two weeks.
Our goal is to continue to reduce wait times, increase your productivity, and help us continue to scale our work from tens of thousands of requests to hundreds of thousands in the years to come. If you have thoughts, feedback, or experience scaling customer support yourself, we’d love to hear from you.
Bulk bug fixes and logging for the DocumentCloud Python wrapper
Thanks to early users, we also diagnosed and hopefully quashed some occasional bugs that cropped up with the DocumentCloud Beta Python library. Issues would occasionally occur in large uploads due to how it handled bulk file handling.
The update also adds logging and error handling to
upload_directory. To use logging, pass a log level of logging.INFO to the client when initialized or configure logging for the
documentcloud logger in your script. To use the error handling, pass
handle_errors=True to the
We’ve also added retry logic to some requests which previously did not have it, so timeouts should be less common overall now. This update also fixed the logic for saving projects to respect the bulk limit when adding projects.
Image via Wikimedia Commons