Today we announced the release of the VAMOS® technology platform. This brand new technology stack represents a significant release for Simulmedia that enables us to take a technical leadership position in scaling attribution-based audience targeting for television with data-driven software solutions. The platform now gives us the ability to tie TV ad exposures to actual conversions including sales, website visits, and other ROI metrics.
VAMOS Technical Overview
Describing the build of the new platform over the last 18 months starts with an initial decision to not reuse a single line of code from legacy software, but rather to take advantage of the ever-changing big data technology landscape to build the platform from the ground up. In other words – new software, new tech, new talent. Fundamentally, we already had robust product requirements and learnings, and needed to expose them in software that did not exist.
The VAMOS stack is entirely cloud-based on AWS and has been architected using micro-services for performance, scale, and flexibility. Our goal has been to create a strong foundation in data ingestion and storage, while also making heavy use of APIs and allowing us to use purposefully-chosen tools in solving unique problems. For example, as part of our target audience creation tool, we used a high-performance key-value cache Redis as a database for generating smaller queries that required near-instantaneous response.
Figure A: VAMOS Marketing Technology Stack
At the core of the data-driven platform is Amazon Redshift, a massively-parallel processing (MPP) data warehouse. With its robust support for SQL, it has become our single source of truth (SSOT) for all teams within the company who want to be able to explore data through a variety of industry-adopted tools. Interactions with the data warehouse are conducted via sets of APIs that are generally designed to solve unique functions.
Our frontend technologies are where we take some liberties given the need to continuously evolve visualization experiences across all different applications. Ruby allowed us the opportunity to build webapps quickly while taking advantage of shorter learning curves and robust community support. Additionally, tools like Ember.js and D3 have allowed us to create rich data visualization experiences delivered back in the form of web-based reporting applications.
Figure B: VAMOS Reporting
VAMOS Platform Features
In order for VAMOS to deliver powerful targeting applications, our platform first aims to solve the problem of making TV data behave and perform like web data. This is very difficult. Unlike the digital world where there are generally well-established standards on data transfer protocols, performance, SLAs, and formats, this is almost non-existent in linear TV. Therefore, our first mile problem of bringing large, disparate, unreliable sets of data into the platform, processing and loading it in the data warehouse, is where a significant amount of Engineering challenges occur – and where a significant amount of our time is spent. Once complete the data is staged in the data warehouse and made available for us in our applications.
VAMOS enables us to build powerful audience targeting applications on a robust data set. Applications provide the ability to move users from the initial planning stages of a campaign to creation of the media plan to gathering campaign reports through a self-service UI. VAMOS is made up of 3 distinct applications:
- VAMOS Audience Engine. This UI allows users to intuitively build demo, purchase, and program-based audience targets for a campaign and expresses them as a percentage of the entire TV viewing universe - in under 5 seconds.
- VAMOS Plan Optimization. This UI guides users in maximizing campaign results by using a robust control interface to intelligently allocate media with full support for building complex media plans. These queries take seconds to run, allowing the user to generate full media plans in 3 - 5 minutes.
- VAMOS Reporting is a self-service portal that takes advantage of robust data visualization frameworks to provide deep insights into campaign performance and market intelligence – along with the ability to quickly receive data in other formats.
Figure C: VAMOS Audience Activation
Most importantly, in bringing transparent ROI to the advertiser, the platform enables us to match viewing data with purchase data to prove the incremental impact that TV has. It’s an exciting time for Simulmedia as we deliver solutions that make TV as accountable as digital.
This 18-month build of VAMOS was the resulting work of 29 members of the Engineering & TechOps teams, 6 members of our Product team, and 8 people in our Data Science group. It goes without saying that in order to build really great software you must have really great teams. We are extremely proud of the tech talent we have at Simulmedia and look forward to bringing you more very soon.
Sanjay from Engineering
Interested in getting the latest from Simulmedia?
News, insights, and events sent straight to your inbox!