It seems everywhere I turn, leaders in Media & Entertainment are catching-on to the Big Data phenomenon. At the very least, there’s a whole lot of conversation happening about the subject. So, it wasn’t surprising that nearly 800 people registered for a webinar – sponsored by world leader in analytics, Teradata - to hear online auction (and analytics) powerhouse, eBay, share how they’re using Teradata and Hadoop to access Big Data sources , and analyzing them to give eBay speed, power and agility to offer their customers what they want, when they want it. (If you’re still catching-up on what Big Data means exactly, read my earlier blog on the subject.
The conversation, led by Teradata’s Chris Twogood and eBay architecture and technology rock-star Tom Fastner, outlined a variety of ways that eBay relies on Big Data analytics to touch almost every part of the customer experience, from driving traffic to the site, delivering the most relevant product assortment while there, and by analyzing every step of the final transaction.
Here are a couple of gems from the webinar:
- Paid Key Word Optimization: I don’t think I was the only one whose jaw dropped when Tom mentioned that eBay manages 167 million paid keywords. To put that in perspective, the English language has just over 1 million words. "Keywords” are often combinations of frequently searched words. In eBay’s world, think of phrases like "iPad 2” or "gold diamond ring.” (In the Media and Entertainment industry, you can imagine combinations like "Tom Cruise action movie” or "Harry Potter Lunch Box”). When bidding on key words – no matter how many you manage – the bid process also includes things like geography , and rank/placement (do you want the keyword which appears first, second or third on the list). So, when all said and done, these permutations multiply.
eBay uses Big Data Analytics to identify those keywords most commonly searched on the eBay site, and which drive the most traffic and conversions, and ultimately decide on how much those words are worth when bidding. Across nearly 200 million words, you can imagine the massive cost savings.
- A/B Testing: So, once eBay brings potential buyers to the site through optimized key words (and other significant means, of course), the goal is for those potential buyers to become actual buyers. eBay shared the fundamental role that experimentation and Big Data analysis play in conducting and determining the outcome of those A/B tests. eBay conducts A/B tests on 100 use-cases at a time, resulting in thousands (tens of thousands) of simultaneous A/B tests. They’re able to measure response to those tests and determine the best product assortment , and layout (e.g. do you prefer bigger or smaller pictures?) to match buyers’ search requests, resulting in better matches and more transactions.
If you think this sounds interesting, but are still left wondering what any of this has got to do with us in Media and Entertainment? The parallels are striking:
- So What About Search? - Driving traffic through search – with more content aggregators and distributors competing for the same audiences, you want to make sure your audience finds you first. Understanding and optimizing what drives traffic – even with keyword tallies infinitely smaller than 167 million words – is critical. Conventional wisdom use to hold that you should always bid for the #1 slot, but smart advertisers know that other slots outperform the #1 slot on an ROI basis, so unless you need the volume the first slot drives you're better off bidding for slots 2 and higher.
- Content is Personal too – No doubt about it. You know that your movies, episodes and trailers are every bit your product, just like an iPad on eBay. And, for most properties, you have merchandise you can pair. Learning what your audience likes, and what resonates with them , through A/B testing to optimize recommendations and product assortments will only increase site engagement, and ultimately, transactions.
Did these gems pique your interest? Then don’t forget to listen to the webinar replay in its entirety.