This is an internal project of Applovin, the task of which is to help growth managers of the company to run effective advertising campaigns: launch ads, settings, analytics, prediction.
LionMachine — complete tool for managing marketing campaigns
At the time when we started the project, the company's growth managers had to use 5 different tools (adjust, facebook meta ads, applovin, box and asana) to do their job correctly. This process slowed down the work plus there was a very high risk for managers to make a mistake, because the data had to be constantly compared in different services. The company decided to merge all these services into one - LionMachine.
Problem & Solution
Applovin hired whole separated team for this project and I was among the first to join it as a product designer. There was also a ui specialist in the team, who later helped with the creation of ui-kit and visual style design. Since there was no product manager in the team, only tech lead, some product tasks (describe solutions, talking to stakeholders and etc.) had to be done by me.
My role
As i mentioned it was internal product, thats why i had access to almost all 50 users for whom this project was being developed. When i started i knew absolutely nothing about how the mobile advertising industry works, so for the first month I played the role of junior growth manager. I participated in meetings with managers teams, tried to set up advertising campaigns, conducted interviews with lead managers, studied various materials on the Internet. I recorded all the meetings, analyzed and wrote out key points: how managers work, what actions they do, what they use more often, what they pay attention to, etc. My goal was to understand how managers think. The work done helped me to form a vision of the project, to break it down into logical parts, and to understand where the work should begin.
Diving into the topic
After I began to understand the inner workings of this market, I began to analyze the tools used by managers and the only potential competitor - Bidalgo. Almost all services i analyzed directly with the managers who use them. We were interested in both the functional component (what features the service has and what it lacks) and the implementation in terms of user-friendliness. This step allowed us to define the functionality of our final product in detail, as well as to group it logically.
Tools and competitor analysis
Since the main task of the product is to conveniently and quickly solve the managers needs, as well as to save development time (think money), we decided to move with MUI framework, which is easy to customize for our needs both in terms of design and development.
It was unnecessary and difficult to offer something special for visual design, because the whole project is about working with a lot of data, and it is easiest to work with them when they are arranged in a table, so my task was to design the elements (filters, buttons, search, settings, etc.) as user-friendly and logically as possible, so that managers on an intuitive level understand how to work with this system.
Prototyping (design) and testing
As I mentioned above, at first I worked directly with managers in the team, studied how they use different products, and noticed that every day they do the same thing: they evaluate the overall performance of the advertising campaign, if any of the indicators does not meet expectations (expected ROI 25%, but the real 20%), the manager goes deeper into researching it to find the reason/place where something went wrong. Such work on average takes from 1 to 3 hours, while the manager goes through the whole hierarchy of the campaign and finds the right indicator, and he has from 10 to 20 games in work, so overall work could take more than week, which leads to losses for the company (lost profits).
AutoAlerts
I had the idea that this process could be optimized with the help of "smart algorithms", which we later called AA (auto alerts). The idea was that the manager, when launching a new advertising campaign, would specify in AA the final indicators he wanted to get, set up the frequency and depth of checking these indicators, and launch the advertisement. In case the metrics were lower than specified, the system did its own analysis of where the failure occurred and sent the manager an alert (web, email, slack), but with specific instructions on where something went wrong. This way, the manager skipped the step of finding the problem and went straight to finding a solution. It was possible to implement such a system only when we could receive all relevant data from other services, so it was decided to postpone it until the launch of the main functionality.
Before starting work on AA, we needed to validate the idea, so I designed the functionality and built clickable prototype. We presented it to 10 top growth managers, got their full approval and comments on the prototype, then final design, development, testing, release.
🤷🏻‍♂️
To see my works
please open website from desktop.
Made on
Tilda