Revolutionizing marketing campaigns with AI
SAS helps dramatically improve productivity and model efficacy.
data processing times
Alliant relies on machine learning to create qualified marketing audiences for its clients
For marketers, identifying qualified prospects is critical to reaching consumers. But many companies don’t have the database or the analytical firepower to compile effective audiences. They end up blasting campaigns to audiences without taking preferences – or likelihood to respond – into account just to keep pace with their high-demand environment.
That’s where Alliant, a data-driven audience company, can vastly change a client’s marketing approach – and results. Whether you want to target millions of prospects for a campaign or purge your first party data of ineffective targets, Alliant uses machine learning algorithms powered by SAS® Viya® to quickly create bespoke marketing audiences from a database of 270 million consumers.
“We had a need for speed, so naturally we called SAS. After migrating to Viya, we now have ample computing power to process gigantic data sets much faster. Bill Adam Senior Vice President of Data and Technology Alliant
Lightning-fast data prep
At Alliant, creating high-performing marketing audiences for clients begins with data management. This is where Bill Adam, Alliant’s Senior Vice President of Data and Technology, enters the story. Adam and his team are responsible for all data assets at Alliant.
“Our data comes in every flavor you can imagine,” Adam says.
Alliant, a cooperative database, integrates data from a large network of marketing partners. As each partner updates its customer data monthly, Adam and his team load it into Hadoop, then use an ETL process to retrieve predictors and pass them into SAS Viya as structured data sets for analytics. Each data set includes anywhere from 10,000 to 14,000 candidate predictors for use in the modeling process.
Before installing SAS Viya, this process was cumbersome. The addition of bountiful data sources, along with a rapidly growing base of cooperative members, was ballooning data volumes from millions to billions of records. The previous data environment simply wasn’t strong enough to handle the new reality. At the same time, clients were requesting faster turnaround times.
“We had a need for speed, so naturally we called SAS,” Adam says. “After migrating to Viya, we now have ample computing power to process gigantic data sets much faster.
“In the world before Viya, say we had a universe of 100 million people for a campaign – we’d have to reduce our workload to 30 million to meet a client service-level agreement,” Adam explains. “In the Viya world, we’re running 10 times faster, so we can segment our entire universe to create better audiences.” On top of this, Alliant is finding cost savings by creating efficiencies in data prep work.
Alliant – Facts & Figures
of consumer transactions aggregated
Predictive modeling with machine learning
Work continues as the Data Science team – led by Malcolm Houtz, Vice President of Data Science – constructs predictive machine learning models to segment and score large data sets into valuable marketing audiences.
Previously, Houtz and his team relied solely on logistic regression to identify good marketing targets. But clients were asking for more and more prospects. And faster. As Houtz explains it, performing more logistic regressions on a data set only reduces the number of prospects for a particular campaign. He needed to run different algorithms to cast a wider net.
SAS Visual Data Mining and Machine Learning allows Alliant to run multiple machine learning algorithms at once. While continuing to include traditional logistic regression, Alliant has transformed its modeling operations by simultaneously applying algorithms such as neural networks, support vector machine, gradient boosting and random forests to the same data set.
Alliant now can produce larger audiences by including every qualified name from each algorithm – or higher-quality audiences by including only the prospects that all five algorithms qualified as top performers. The company also can apply algorithms specifically targeted to meet clients KPIs, whether that’s response rate or customer lifetime value.
“SAS helps Alliant deliver models in a quarter of the time of traditional workflows and shorten processing times by 85%.” Malcolm Houtz Vice President of Data Science Alliant
A better customer experience
Alliant accesses a growing library of more than 1,200 models to provide audiences for marketers every day – in any channel – from direct mail and email to programmatic and addressable TV. These models are on demand or custom built for clients across industries to help optimize campaigns, provide real-time scoring of data sets, or drive brand awareness.
“The speed is incredible for this volume of data,” Houtz notes. “SAS helps Alliant deliver models in a quarter of the time of traditional workflows and shorten processing times by 85%. This boosts productivity, increases client engagement and generates more income.”
Houtz shares one example where a large client needed a fast turnaround-time on scoring 200 million prospects. “We couldn’t have done it previously,” Houtz says. “But with SAS Viya and our machine learning algorithms enabled with Viya, we were able to turn that project around in one day. The client was delighted, and that’s very good for our business.”
Keeping a competitive edge
“The data and marketing technology industry has evolved at a rapid pace,” Houtz says. “So we have to deliver stronger and stronger models, and SAS helps us do that.”
The implementation of SAS Viya and products has helped Alliant stay on the forefront of industry change, developing faster processing and scalable architecture, enabling Alliant to keep a competitive edge. And, perhaps most importantly, these updates also have provided the structure to prepare and manage data assets to meet and exceed new data compliance requirements, such as the California Consumer Privacy Act.
The results illustrated in this article are specific to the particular situations, business models, data input, and computing environments described herein. Each SAS customer’s experience is unique based on business and technical variables and all statements must be considered non-typical. Actual savings, results, and performance characteristics will vary depending on individual customer configurations and conditions. SAS does not guarantee or represent that every customer will achieve similar results. The only warranties for SAS products and services are those that are set forth in the express warranty statements in the written agreement for such products and services. Nothing herein should be construed as constituting an additional warranty. Customers have shared their successes with SAS as part of an agreed-upon contractual exchange or project success summarization following a successful implementation of SAS software. Brand and product names are trademarks of their respective companies.