Why Big Data Is The Future Of Bionic SEM
Once and for all, I will prove why Big Data is the future of bionic SEM. But first, indulge me for a minute.
In the classic 1970s TV show, The Six Million Dollar Man, when all-American test pilot Steve Austin (played by Lee Majors) teeters on the verge of life after a crash, the government famously declares, “We have the technology to rebuild this man.” By surgically altering him with futuristic bionic technology, Austin is reborn with superhuman strength, speed, hearing, eyesight, etc., and becomes a superhero that could weekly battle (in one episode) bad guys, aliens, and even Bigfoot.
The famous tagline from that show was “Bigger. Faster. Stronger.” – a reference to the upgrades and improvement made to Austin.
I can imagine there are a lot of Paid Search pros reading this post that would also love to bring some bionic technology into their enterprise-level accounts and reap the benefits. The good news is that it’s already here. The one thing that definitely has gotten bigger, faster, and stronger is data. It’s gotten bigger in volume, faster in delivery, and stronger in benefits if used correctly.
I recently spoke at SMX Advanced in Seattle on this theme in a paid search session entitled Big Data + Big Math = Exponential Search Performance. Along with data leads from IBM, Microsoft, and Majestic SEO, I presented on how Big Data is literally changing the game.
Using the latest engine solutions and innovative technologies from trusted vendors (such as my own Kenshoo), Paid search marketers can take their campaigns to new heights that they could never reach before. Innovative, multi-touch attribution measurement, machine-learning, algorithmic bidding engines and advanced targeting are just a few of the bionic options that Big Data is bringing to SEM.
Check out this page on several great working definitions of Big Data. My current favorite is: “Big data is what happened when the cost of storing information became less than the cost of making the decision to throw it away.” — Tim O’Reilly.
Still Not On The Big Data Bandwagon?
Because Big Data has become a virtual buzz word in this industry, there can be some eye-rolling accompanying its use. In my presentation at SMX, I tried to hammer home the fact that no matter what your current job function is, everyone in this industry – from dedicated, number-crunching analysts to client-facing account managers to even the CMO – needs to get comfortable with Big Data.
In a channel such as SEM, which has become the science of squeezing every bit of efficiency from an account, Big Data can ultimately contribute to the success of your company or allow a competitor to leapfrog you in the market. As I reflect upon the core elements of paid search (keywords, creative, management, reporting, optimization and integration), I can see ways that search marketers can use the size, speed, and strength of Big Data now and in the future.
Let’s Just Examine Big Data With Keywords
Is there any data set marketers touch that is more granular than keyword-level data? In some sense, keyword-level data is the original Big Data in this industry.
An average enterprise-level account may easily contain tens of thousands, hundreds of thousands, or even millions of terms that can quickly fill the 1,048,576 rows by 16,384 columns in Excel. Yet, I think most search marketers would agree that they would feel very handcuffed if they were forced to work at the ad group level, which is just one degree of granularity above the keyword.
Well, the most granular level is not the keyword level, but rather the keyword level x attributes such as match type, geo-location, device type, ad format, etc.
Let’s say you have just 10,000 keywords in your account. A pull by match type would triple or even quadruple (if you use modified broad) to 40,000. By device? Just desktop, smartphone, and tablet would triple it to 120,000. By engine with just Google and Yahoo! Bing Network would double it to 240,000. If you just wanted it by state (not even DMA or city or country), that would be fifty times greater and well over 10 million rows.
Now, what if you wanted that by day for a 90-day campaign? Almost a billion rows of data! What about by each of the twenty-four hours?
That’s 24 billion rows of data from 10,000 keywords…
Of course, you could dispute that the level of data I’ve described is too granular and not applicable to paid search optimization. Really? Would anyone like to argue that there’s no difference between a click for tennis shoes on exact match, a tablet, on Google, and in Florida versus tennis shoes on broad match, a desktop, on YBN, in Minnesota?
What about a Sunday morning versus Friday night? A search text listing versus a PLA?
Still not worth checking out? You would absolutely bid the same, use the same ad copy, and apply the same budget for each of those very simple examples? Of course not!
Now, imagine you were the only company in your vertical that could pull out insights that deep. You can be — right now, in fact. What are you waiting for? Your competitors will get their act together soon. By then, you’ll be on to Bigger Data, Massive Data, or even Gigantic Colossal Data and stay one step ahead. C’mon folks, let’s get bionic!
Still rolling your eyes at Big Data?
Big Data & The Other SEM Elements
Following are some thoughts on how Big Data is changing the game for the rest of paid search.
Reporting & Management. Big Data is requiring next-generation tools to handle the complexities involved in these very large data sets. For those of you who haven’t yet experienced the real-world concerns when your data can’t fit into Excel, there are many issues. You may have a big appetite for Big Data, but if your organization doesn’t have the server or processing power to manage it, you may as well be working at the ad group level.
And once again, where there is crisis, there is opportunity. Not every company is going to be data-mature enough to handle the logistics of Big Data. If you can get your company there quickly, you could be the only company in your vertical that is able to harness it.
Creative. Dynamic keyword insertion (DKI) has been available for years and has been proven to be a fairly effective tactic when used strategically. More advanced tools are available in this same vein.
Not to toot my own horn here, but I personally think our (Kenshoo’s) real-time campaign (RTC) functionality is pretty nifty. It can pull in your online store’s product feed and create ads from combinations of your product attributes.
In the example below, the RTC ad combines the brand and the product into the headline and then displays the inventory left as a way to urge customers to purchase. It can even detect when you’re out of stock and pull that ad so you don’t pay for ads that you can’t convert.
What’s better than Big Data? Automated Big Data!
Optimization. Optimization is the lifeblood of paid search. It’s why this industry has grown so quickly from millions to billions of dollars in spend and many times that in return to advertisers. We are able to learn, adapt, and bid accordingly. As the data exhaust keeps growing larger from our campaigns, many of the tried-and-true bidding methods and models will become extinct. New, evolved bid policies with machine-learning capabilities and driven by algorithms (ones that even Neo from The Matrix couldn’t comprehend) are needed.
Integration. We’ve known for years that search does not exist in a bubble. If anything, search is one of the most complementary advertising channels in the marketer’s toolbox. Integration is the great new frontier for SEM – and it can only happen in a connected, Big Data world. Search + Social in itself is a compelling and intriguing concept. Search + the rest of the channels (even offline) is a true game changer for search engine marketing.
Big Data is as Big Data does. Anyone who tries to downplay the value of Big Data instantly has zero credibility in my book. Do we have a ton to figure out? Yes. Will we have all of the answers soon? No. I’m not saying that you should dive in at full speed tomorrow. I just think that if you’ve been completely ignoring the potential already being generated by the data from your search campaigns, you may want to start chipping away at the problem… before it becomes bionic.
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.
(Some images used under license from Shutterstock.com.)
No fluff - just the best news in paid search marketing every week.