The goal of this project was to create a unique analytical system with an enormous database that eliminates the need for a deep research and gives clients the best recommendations on their ad campaigns within seconds.
Up to 250%
campaigns' setup speed increase reached.
Over 1.5 Billion
ad impressions worldwide scored.
Together with Meme Global, we developed a unique platform that helps their clients to create and launch the most effective ad campaigns. Earlier, it took about two weeks to setup one video ad campaign, now the entire process takes only a few minutes. The platform gives the most precise recommendations on targeting, location and time, which is the result of our deep video ad market research, careful technologies selection, and the implementation of our specially designed cutting-edge analytical tools.
This project intended to solve two main problems. On the one hand, we needed to eliminate the necessity in spending enormous resources on research to define precise target consumers, prime-time and a lot more parameters to build and set up an effective ad campaign. On the other hand, we needed to automate a complicated work of the analytical team, so they won't need to constantly hand-check all the information and parameters to give customers suggestions required for their future ad campaigns. By automating the process of data gathering and analyzing we made the new system generate the most effective and precise recommendations. Also, the system had to be easy to navigate for an average user.
From the very beginning of working on this project, we provided a variety of researches to define what should be used for each stage. Thus, we needed to find out what exact tools, methods and technologies to apply for data gathering and mining. The main difficulty lied in choosing technologies that would interact with RTB platforms and fit for creating Ad Servers and CRM system perfectly. The test versions of the system's component parts were written using PHP, Python, and Node.js to see what technology suits most for creating the final product. It was decided to choose Python as it has the resource to make the system scalable and flexible to adding new advertising platforms (i.e. search, social, display ads, content advertising and more) in addition to the existing video ad platform.
In terms of this stage of working on the project we had to make a broad-scale video ad market research. This research helped us to create special tools for automating the research entirely. Basing on the information gathered from thousands of websites and subject directories we created ad placement analysis flow and parameters for scraping bots. These scrappers had to track information from different locations and devices at different times of day to collect all possible information on video ads (providers, final consumers, etc.). It required 10 servers to ensure non-stop data collecting, processing, and renewal. At the same time, we used Silenium to emulate profiles and actions of real users to ensure the smooth work of the platform. After gathering all information needed, we delivered it to analysts for further processing and creating the Brief Buster feature.
The core of this challenge lied in the necessity to create a unique system that would analyze a huge amount of information, generate the most precise and effective recommendations for each client, and give them the ability to adjust their own ad campaigns and set them to the maximum efficiency. Our collaboration resulted in the creation of the Brief Buster feature. In brief, it is a popup that shows users how leading global companies adjusted their ad campaigns in order to use this experience for their own video ad campaigns.
Shaping all collected data into one integrated information database was highly challenging yet achievable. It was impossible to place this amount of data on one single server, so we had to use different sharding technologies to achieve maximum database reply speed on requests from users. Also, we decided to combine SQL and NoSQL databases to ensure the highest service productivity. The second main problem in handling this amount of data was the proper database maintenance and support as the information constantly changes in its content and size. We solved this problem by integrating different API services that continuously provide support and keep information on ads from the variety of advertising platforms updated.
The last but not least challenge was to gather all managerial and other feature sets, data warehouses, and analytical subsystems into one wholesome CRM system that would increase the client's business efficiency. As a result, the system gives managers the ability to assist clients with their ad campaigns, and an access to clients' actions and payment history. In addition, the clients can use an easy-to-navigate web application that offers advanced creating, customization and navigation of their advertising campaigns.
Together with Meme Global we created a unique system that makes the process of market research and an ad campaign setup fully automated and as fast as it can be today. This project was a huge challenge for us, yet it became a reason to be proud of ourselves. Months of our deep research, data analysis and software development have reduced the time spent by the Meme Global clients on adjusting and setting up advertising campaigns to 11 minutes. Today, any client can build their campaign 250% faster basing on the experience of leading worldwide companies, and increase their success in business within seconds.