How Convermax Team Handles BFCM
Black Friday and Cyber Monday are the most critical days of the year for any online store. As a search service provider, we understand that during BFCM, our customers rely on Convermax to operate flawlessly, without slowdowns, surprises, or downtime. We want our customers to make as many sales as possible without anything disrupting the process. That’s why we treat BFCM as a highly complex operational event.
How we prepared
Extended team working hours
For the entire BFCM weekend, we ran a special support and engineering schedule. Half of our support team was actively monitoring tickets and store performance on Saturday and Sunday from 8:00 am to 4:00 pm EST. Several engineers from the product development team, including our Chief Engineer, were on call throughout the weekend during the same hours. If some major problems could occur, they were ready to step in and fix them.
Infrastructure scaling
Ahead of the sale, we added additional servers to handle traffic spikes. Our goal was to ensure that increased traffic would not impact search speed or overall stability for our customers.
Convermax is built with a clear separation between indexing and search. This means indexing and search infrastructures run independently. Even if the indexing goes down, shoppers can still search products as usual, as if nothing happened.
Code freeze
In the week leading up to BFCM, we enforce a full code freeze on all global changes. During this period, we do not deploy any updates that could affect all stores at once or impact our server infrastructure. Small, store-specific tweaks were still allowed, but only after thorough review and when they affected a single client.
Fast response and proactive monitoring
It’s not our first BFCM weekend. Since 2011, there has been no BFCM-related issue that took the Convermax team more than half an hour to close. Most of the critical issues are resolved within 10–30 minutes of being reported.
Our system continuously monitors API health. In addition, we run automated UI tests every few hours to catch visual or functional regressions early. These results are currently shared in our internal Discord channels so the team can react quickly.
BFCM results
Everything went smoothly. We had only two minor issues reported during the entire BFCM period:
One issue was caused by a slightly outdated index and was resolved after pushing a store re-indexing.
The second was caused by a third-party review service provider’s API going down, which we fixed by temporarily turning it off.
All issues were resolved within half an hour.