Early Birds Seamlessly Absorbs Increases of 50–300% in its Web Traffic
The data flow processed by Early Birds can triple instantly. This happens particularly at the beginning of sale seasons or when a large group such as Fnac or Darty becomes one of the company's clients. As a result, Google Cloud Platform was the obvious choice for Early Birds, which specializes in real-time personalization of customer journeys.
Products Tailored to Each Internet User
To predict expected consumer behavior, Early Birds uses algorithms that are carefully developed by the company itself, its clients, or third parties.
"We help major retailers to increase the conversion rate of their e-commerce sites by between 3% and 15%," says Nicolas Mathon, co-founder and Associate Director.
In order to do this, Early Birds analyzes a range of data, including: clickstream data provided by retailers (the pages visited, time spent on a product page, and any subsequent actions); what the Internet user is interested in (frequently visited sections, notifications, wish lists, and data from social networks); the conversion funnel (add-to-cart rates, purchases, returns, abandoned carts, etc.); and in-store purchases and telephone orders.
With Early Birds' Application Program Interfaces (APIs), merchandisers and product managers can customize up to 90% of the content on their retail website, tailoring it to the type of consumers currently browsing the site. Early Birds' recommendation engine dynamically determines relevant recommendations, depending on the consumers' tastes and preferences, and places them in defined areas in accordance with the brands.
"Thanks to Google Cloud Platform, we process an average of 600 predictions per second at any given time, reaching up to 2,000 per second in sale seasons. We optimize the potential of customer data in real time," says Nicolas Mathon.
From Virtual Servers to Managed Services and Predictive Analytics
Early Birds has been a Google Cloud Platform customer since 2013 and originally used Google Compute Engine to test its algorithms and analyze its consumers' histories, Google Cloud Storage to complete its automatic backups, Google Cloud CDN to display personalized content relevant to the consumer, on a country-by-country basis and in near real-time, and Google Cloud Load Balancing to optimize load distribution between the Hadoop/Spark virtual servers.
"Our technical infrastructure has been hosted by Google for several years. We never would have been able to expand at this pace with our own servers," explains the co-founder.
In 2016, Early Birds moved up a gear by using Google Cloud Platform's managed services. "10% of our employees used to work on installation, administration, and maintenance tasks. We decided to outsource these functions so we could concentrate more on our area of expertise: algorithms," he adds. For developments, Early Birds uses Google Cloud Dataproc and Google Cloud Dataflow managed services, activating just a few servers or dozens, on demand, depending on the complexity of the algorithm tested.
Alongside managed services, Early Birds completely overhauled its decision architecture in 2016. The company transferred all of its databases to Google Bigtable and began using Google BigQuery for its predictive calculations. "Our SQL queries were running too slowly for the amount of data we were processing. Now, our experts can simulate conversion funnels in just a few seconds, enabling our customers to benefit from this valuable information."
Google Cloud Pub/Sub is now at the heart of the API management platform, recording all interactions and events triggered by consumers. Moreover, thanks to Pub/Sub, Early Birds can now offer its customers the opportunity to develop their own APIs within the company's technical environment (framework).
"We have not found anything that can rival Google Cloud Platform and match it in terms of performance or reliability. Google knows how to mask the complexity of its technologies. We are making the most of its power while retaining transparent services and a reliable infrastructure. Google is also a huge innovator in terms of algorithms and artificial intelligence. We are interested in its new Cloud Spanner technology for our future international expansion. It will make it easier to manage our APIs across the world," concludes Nicolas Mathon.