Black Friday & Cyber Monday 2025: More Hardware, Less Latency

As the dust settles on another record-breaking Black Friday and Cyber Monday (BFCM) for GBG Loqate, it’s clear that preparation and new infrastructure were the keys to success in 2025, achieving a huge 30% reduction in platform latency. We have gone from 6 billion annual requests in 2016 to over 1 billion just on BFCM alone in 2025. Building on last year’s learnings, we doubled down on resilience, teamwork, and the underlying hardware because when an estimated 41 billion USD is spent over one weekend online, every millisecond matters. 

Another Record-Breaking Year for GBG Loqate

This year’s Black Friday/Cyber Monday period was our busiest yet, with traffic to our cloud platform surging more than 6% compared to last year. But while the growth in demand was exciting, the real headline is how our infrastructure performed under pressure.

We achieved a 30% reduction in latency, delivering an average response time of less than 55ms throughout the weekend. To put that into perspective, our premium SLA guarantees 350ms so maintaining less than 55ms during the busiest shopping period of the year is a milestone we’re incredibly proud of.

This achievement reflects the strength of our engineering investments and the resilience of our platform. As traffic continues to climb year over year, we’re committed to ensuring that speed, reliability, and customer experience scale right alongside it.

So, what did we do differently this year?

1. Request Caching for Spam Protection
2. Risk Management & Bespoke Monitoring
3. New Data Centre
4. Collaboration & Change Management
   
   

Let’s have a look at the above in more detail:

1. Caching & Spam Protection

Improved Front-end caching became a hero feature in 2025. It helped us fend off spam requests that could have otherwise slowed down legitimate transactions. Among the spam we encountered this year? A CV for a job application at a popular UK supermarket chain masquerading as a capture address verification request, proof that you never know what will hit your API’s!

2. Risk Management & Bespoke Monitoring

Risk was front and centre this year. We looked at the data from last year and identified customers that followed certain traffic trends that may impact latency. This allowed us to create the ability to quickly make changes to certain backend systems for those customers if required to mitigate any issues.

We also introduced more bespoke monitoring, giving stakeholders granular insights into error rates, request volumes, and year-on-year comparisons. This transparency wasn’t just reassuring; it empowered smarter decisions in real time.

3. Our New Data Centre

If you read my previous article Black Friday Brilliance: Managing a Billion Transactions Through Tactics & Teamwork, you will have noticed me talk about our plans to bring a new data-centre online in the USA. Well, this year we managed to get that infrastructure up and running before BFCM. This was a risky move, only having done so a few months prior to BFCM but the benefits outweighed the risk from our point of view.

The risk paid off! Our Premium SLA promises 350ms, with some customers constantly monitoring our response times for any breaches. With the extra capacity and newer underlying hardware from our new data centre, we managed to not only keep to our SLA but outperform it by nearly 300ms over our busiest event of the year. This kept our customers very happy and support team stress free!

4. Collaboration & Change Management

November remained a low-risk change zone, with only low risk changes being merged into the products, and a complete code freeze in the second week to lock stability ahead of the big weekend. This was slightly earlier than previous years as the risk was higher due to the new data centre coming online a few weeks prior. Weekly meetings with a dedicated BFCM core team ensured alignment across engineering, support, and operations. We also extended customer support meaning we were on hand should our customers need us - although I am pleased to say, we had zero reported issues.

Monthly Load Testing Continued

Last year, we introduced monthly load testing to simulate peak traffic scenarios well ahead of November. By doing so, we identified bottlenecks early and optimised performance under pressure. This proactive approach meant fewer surprises and smoother sailing when the real surge hit.

Conclusion

BFCM isn’t just about surviving the surge for us, it’s about thriving under pressure and making events like this become business as usual. We have come a long way since 2016 with 9x more requests.

Whilst on the day everything went smoothly, some might consider this now business as usual, there was still 2 months of preparation that went into making everything work faultlessly over the weekend.

It is probably not a surprise to you that more and better hardware leads to better performance. This coupled with innovations like smarter caching and better geographically distributed infrastructure, means we’re setting the stage for an even more seamless experience in 2026.

To find out more detail about the preparation for BFCM at Loqate, check out this other article: Black Friday Brilliance - Managing 1bn transactions through teamwork | Loqate | GB