The Situation
It’s not often that you get a small family business interested in something as monumentally mind blowing as RTB arbitrage.
Well, that is unless the family happened to be a duo of experienced Ad Tech professionals. We were approached to build a first proof of concept and an RTB scalable architecture.
This multifeed platform had to unite the numerous supply side platforms(SSP) and demand side platforms(DSP) feeds the client supplied through existing partnerships – 15-20 unique & proprietary FTP integrations. Many of which were Tier 2 & 3 markets meaning low latency was the top priority in making this RTB business model financially successful.
The Solution
Fig 1 : In a flowchart, everything looks so much simpler
At the time(2016), we were a relatively new company.
As Tier 2 & 3 markets offer extremely slim profit margins, we understood that low latency was the key to profit.
But, the required complexity created by the integrations risked out software being the bottleneck.
We are lucky in our native Saint Petersburg, there are many deeply knowledgeable experts available to connect and collaborate with. This model of growth has since helped us build an efficient team with deep experience despite our company being only 3 years old.
Scalable architecture – problem solved. We got a guy for that. And thanks to this insight we produced a platform capable of managing 30,00+ QPS.
Load balancer – hmmm…
This we optimized empirically. A lot of trial and error ensued given the number traffic sources for redirecting and matching. We managed this by first focussing on filtering out bots with obvious redirect signals. And then… the rest is a closely guarded secret. But let’s just say it worked out well enough on the proof of concept for a fully viable MVP to be requested.
We went on to build monitoring, reporting and campaign management tools.
The Success
After 6 months, the MVP was in action, an end to end RTB arbitrage system including: load balancer, monitoring tools, reporting & campaign management components. It’s been running smoothly ever since.
We’ve continued to work together and build in additional features at the client’s request. They’re enjoying family life and an ongoing income in addition to initial 3-fold ROI!
Our lesson learnt at this early stage was that having experts on hand to avoid pitfalls and costly redesigns really reassured our client that their upfront investment was in good hands.
It became standard practice for us to optimize and re-optimize workflows so much that we now automate our QA process. This really pays off in our clients’ bottom line!
Feedback
Source: GoodFirms.co
Programming Stack
J2SE, Servlet API, Spring beans, Kafka, Redis, PostgreSQL, Zabbix.
Can we help you?
Maxilect does more! Get in touch to find out how we can build solutions for your digital needs. Our team stands behind all products we designed or improved – we’re there start to finish and beyond.
Call for a free consultation: +7 (812) 775-60-86
Drop us an email: sales@maxilect.com