Procter & Gamble uses Teradata cloud analytics for global marketing

Procter & Gamble (P&G) is the world's second largest consumer goods firm, with 25 brands and more than $84 billion (£52.18 billion) in sales for 2013 to its name. But its global reach and familiarity with consumers has not made it immune to the disruptive power of the worldwide web.

For marketing, it once relied largely on one powerful medium: television. Now consumers spend more time online, through smartphones, tablets and PCs, and the old medium is losing some of its potency. While P&G has invested in online advertising and created around 15,000 consumer-facing websites, it sells almost exclusively through retailers. This makes gathering intelligence on customers a tough job. Getting an understanding of consumers in the online world needed a new approach, says Tony Hudnell, associate director at P&G.

For example, marketing teams for P&G's global nappy brand Pampers have created Pampers Village, a website that offers advice, free samples and money-saving coupons tailored to each stage of a woman's pregnancy or the baby's age, which users select with a graphic slider at the top of the page. In return, consumers have to provide personal details as they log in to the site.

With many P&G brands creating similar initiatives, the firm had to decide what to do with data collected from thousands of websites around the world. 

Dennis Devine, associate director, consumer solutions, P&G, says: “We had huge amounts of data and it was all over the place. There were no standards for it, there was no way to view it. We took it on ourselves to set up a vision of how we could bring all this information together.”

Customer-centric architecture

The global business support centre in charge of consumer relationship management started to build a conceptual architecture, which placed the consumer at the centre, and listed each possible 'touch point' around them, says Hudnell, who oversees all global technical development related to consumer relationship management.

A touch point might be online, in-store or on television. Each was codified and built into an overarching view of consumer interaction. Prior to this, each brand put itself at the centre of the model, Hudnell says.

Within consumer websites, P&G began to collect data for all named consumers into a single repository, including all the choices about how they want to interact with brands and any programmes they participate in.

P&G selected Teradata to build and host a data warehouse, campaign management and analytics system in the cloud, on a pay-per-usage basis.

“We can look at demographics, which consumers are engaged with a marketing programme, what are they engaged with, and what is driving them. We have information on transactions, can see who is most loyal, and see how we can make other consumers more loyal. All these questions we can answer with this tool,” says Devine, who presented with Hudnell at Teradata Partners 2013 in October.

The tools also demonstrate how effective marketing campaigns are and help design more effective campaigns in the future, Devine says. “Creative agencies were saying 'this is the best way to do it'. We did not have visibility into what was really working. Now we have, and we have reviews with the business and they say they can understand what is working and not working.

“This is all very important information to help steer the business. We were basically blind to it before,” he says.

We have data on transactions, can see who is most loyal, and see how we can make other consumers more loyal

P&G uses Secure File Transfer Protocol to extract transfer and load data to its data warehouse. Master data management is supported by software from Trillium.

Teradata launched its data warehousing and analytics technologies in the cloud last month with some fanfare, with P&G as the lead customer. The cloud-based system is built on a Teradata data warehouse, incorporating Teradata Customer Interaction Manager, the Retail Logical Data Model, Teradata Master Data Management and analytics tools including Teradata Warehouse Miner.

Devine says: “We did not want to buy hardware and cobble all this together. We said, 'you're the experts, you can do this for us’, so we can focus on strategy and the business results. This was a new venture for [the supplier], we've had some learnings along the way, but we're really flying now.”

Hudnell says the “magic” in getting the new system up and running was not in the technology, but in the original conceptual architecture which codified consumer actions and determined the data model.

Getting buy-in from the business was also straight forward, he says. “This was not a hard sell, but it was hard to do,” Hudnell says.

“The system is now wonderful, but this was a long journey. When we started in Asia, neither Teradata nor we knew how to do this. We learned together. We loaded data multiple times, we did all kind of things to try and figure this out, because we were going into a space nobody had touched."

Asia and the Americas are now live, and western Europe is about to go so, he confirms.

Proponents of the cloud computing model promote its virtues of speed of deployment and low capital cost. In the case of data warehousing, P&G's experience shows much of the hard work lies in defining the conceptual architecture and data model, whether you use the cloud or not.