It’s no secret that technology is changing the way we shop. More and more people are shopping online than ever before, with over 36% of UK sales being made on a tablet or smartphone according to research from IMRG and Capgemini. 

Unfortunately around 96% of shoppers visit websites and leave without buying or booking anything. E-commerce brands are constantly looking for ways to turn these visits into conversions. Through ‘dynamic personalised retargeting’ brands can engage users by customising content and banners in real time, whether on desktop, mobile devices or social media. Programmatic ad solutions provider myThings uses data-driven programs to help its clients increase clicks and conversions. 

Data is a crucial component of being able to target users effectively for brands such as Adidas, Microsoft and Walmart. The insights gained from data analysis informs the tailored retargeting campaigns that myThings offers to its customers. Therefore, having the ability to collect and analyse that data easily – and then turn it into actionable insights that feed back into the business – is crucial. 

Challenge

MyThings takes in 20 terabytes of data daily – that’s the equivalent of 40,000 hours of television downloads – and customises over five billion impressions. Due to this huge volume of data, myThings needed a reliable and powerful solution that could cope with this sheer amount of data effectively. The engine had to be able to make calculations from a massive volume of current users and also deal with a significant growth in traffic in the future. 

The tremendous growth in data that myThings was experiencing was pushing its systems to the limit, and with the increasing importance of being able to analyse information in real time, anything that took over 12 hours to process was far too slow for today’s marketplace. There was an urgent need to reduce data analysis and processing time to as little as possible. 

Solution

With these high data demands, EXASOL–  a high performance MPP (massively parallel processing) in-memory analytic database – could cope with the real-time pressures of delivering the insights required by myThings to deliver the best programmatic ad targeting solution to the marketplace. 

Most importantly, it is flexible enough to grow with the ever-expanding company and its customers’ data originated in web/mobile/social platforms. 

Ofer Ohana, business intelligence team leader at myThings, said: “MyThings is expanding as a business and we have some high-growth targets over the coming years. 

“We need a solution that is flexible and can grow with us. EXASOL has that key maximum scalability feature. It can keep up with the rapid growth in consumer data in the long term.”

Due to the optimal scalability of EXASOL, myThings was able to upgrade their servers multiple times in just two months – all without incurring any extra costs.

MyThings has been able to grow its system rapidly from four to six servers with a total of 4.5 terabytes of memory. This allows the company to run complex analytic workloads on more than 20 terabytes of raw data. 

Process

Ofer and his department evaluated many solutions to fulfil their real-time analytics needs – including OLAP cubes, Amazon Elastic Map Reduce, Google BigQuery, and Amazon Redshift – but none of these solutions addressed all requirements. In particular, they found that the NoSQL alternatives were lacking the kind of workflow features you would expect from a product like EXASOL.

During a rather quick proof-of-concept process, which was conducted on a cloud installation with a 4 terabyte disk and 255 gigabytes of memory, myThings noted the following strengths of EXASOL that particularly stood out:

•    Fast, user-friendly, and very flexible professional tool
•    Low maintenance
•    Smooth extract, transform, and load (ETL) process
•    Ability to connect to R and SQL Server

Results

In today’s digital marketplace, conducting and acting upon data analysis in real time can make or break a company. If analytical technology is not applied in the right way, brands risk being left behind and will struggle to remain competitive. By identifying and surfacing data insights quickly, and acting upon it in real time, myThings improves its competitiveness and the performance that it is able to deliver to its customers. 

With EXASOL, myThings is able to apply complex data science methodologies to analyse and determine audiences’ interests and affinities with a truly relevant, personalised approach. This approach increases the quality and relevance of advertising generated in real-time, maximizing campaign performance. 

Furthermore, myThings can now reduce query times for each consumer quite considerably. Basket analysis, performance monitoring and reporting are now available in real-time.

Looking forward, as myThings continues on its path of rapid growth, the scalability of EXASOL also makes additional future projects possible. Through investment in the capability to process and perform granular analyses of 100 terabytes of raw data, myThings will be able to save long-term historical data for extended periods of time, thus gaining a truly holistic understanding of each user in its system.