Today’s guest blog comes from Vincent Heuschling, founder and CEO of Affini-Tech, a creator of data platforms that help businesses make data-driven decisions. Affini-Tech is based in Meudon, France and was founded in 2003.
Affini-Tech’s primary goal is helping our customers make data-driven decisions, regardless of the industry they work in. This means giving them easy workflows and web-based interfaces for analyzing and managing Big Data – all at a cost that makes sense for their businesses. Google Cloud Platform has become the foundation for everything we do.
When we launched Affini-Tech, we knew we needed a scalable solution for building applications and analyzing information. We explored a number of Cloud vendors – including doing an initial storage deployment on another large public cloud. However, after trying Cloud Platform and speaking with the Cloud Platform team, we decided to go all-in with Google. Google’s pricing was the most competitive, but we also found it to be the best platform for our developers. Google App Engine, Google Compute Engine and Google BigQuery provided us with an integrated technology stack that worked better than anything else on the market, making it easier for us to build complex applications.
We use App Engine to build applications that help our customers to control their data collections, model data sets and filter and group data. We sell these applications to marketing companies that want to run their software on top of our platform. App Engine is flexible enough to allow developers at marketing companies to customize our stack for their own needs.
Cloud Platform also helps us generate data findings at a faster rate. We’re storing data in Google Cloud Storage, creating ephemeral Hadoop and Apache Spark clusters, then pushing the data into BigQuery for analysis. Ephemeral clusters provide a more efficient, flexible and cost-effective processing model than old-fashioned static clusters and take full advantage of the Cloud model of computing. The key enablers to using the ephemeral clusters are the Google Cloud Storage connector, which lets us directly access data on Cloud Storage using standard Hadoop interfaces, and bdutil, which helps us automate cluster deployment. Our customers only have to “pay as they process,” which saves them money. Not to mention, setup takes less time. Traditional clusters can take days to install, whereas we can get ephemeral clusters up and running in just minutes.
We can pass these cost savings onto our customers, which makes our products and services more competitive. Many of our users are used to spending more than $250,000 to build a data analytics platform. We can often provide the same service for $2,000 per month. This saves our customers money and allows them to go deeper with their data analytics. This access allows our customers to create things like micro segments in their customer base so they can do better targeting for their marketing campaigns.
In a way, Google is helping to democratize data, since more businesses can afford to study it. If a customer is already using Google Apps – and many of them are – we can integrate our data platforms into Google Apps, making these tools even easier to use and understand.
As a small company, the support we receive from the Cloud Platform team is helping us think bigger. It enables us to build new tools and platforms that take advantage of big data. We plan to make a push for business beyond France and the retail sector – and we’re confident about our expansion, with Cloud Platform doing the heavy lifting.
- Contributed by Vincent Heuschling, founder and CEO of Affini-Tech