Big Data, it is a term of which many marketers still don’t know how to act on it. Maybe because they don’t understand it or maybe because it is ‘too big’ for them. But they will need to understand and get a grip on it to not fall behind and in the end get lost entirely.
Earlier today I spoke to Bryan Eisenberg when he was on his way to his keynote presentation about big data at the PPC Hero conference. Bryan, on his birthday actually, made an interesting point which he will also be making in his keynote speech: we might be heading into a direction in which marketers are (partly) being replaced by automation.
Let’s elaborate a bit on that. The idea is based on several different elements which Bryan and his brother Jeffrey describe in a new whitepaper which can be gotten from the website http://www.usethedata.com. It is about understanding big data, or as Rod A. Smith, an IBM technical fellow and VP for emerging Internet technologies says in the whitepaper:
“Big Data is really about new uses and new insights, not so much the data itself,”
Search engines know their main customer is the searcher, which means they have to get the right information to the searcher, both the ads as the actual search results. To give them the best possible results they need to know the searchers intent.
To get to that intent, to understand it, the search engines are collecting data, big data. They want to know as much as possible from the searchers to find their intent.
This means search engines will go and rank results based on personalised data, the data will determine the quality score of specific pages, whether or not the pages are ‘fit’ for the users.
There are several signals which will determine the position of a page in the rankings, both paid and organic.
Until recently the only, or at least most important signal the search engines had was click through rate. That is now changing.
In article on Searchenginewatch John Gagnon a couple of weeks ago shared some interesting information. Two things stood out for Bryan:
1. Bing does not calculate Quality Score in real time
2. The Quality Score is also calculated based on the actual search query, but even more: for Google the search, the landing page and in some cases the conversion also determines the Quality Score.
Using universal analytics together with this data will give the search engines so much data they will be able to provide better answers and thus a better Quality Score.
Marketers will have to adapt to this principle and many will have problems with that, they need to think about this but also need to understand and be able to implement.
Let’s take an example Bryan gave me:
Suppose you have a webshop selling dresses. Now shops might have a page ‘dresses’. But chances are they will maybe have a page ‘black dresses’ (but most won’t), but chances are slim they will also have a page for ‘black a line dresses’.
If someone is searching for “black aline dresses” there therefore won’t be a page available which has the direct answer.
One of the ‘problems’ the search engine will have to fix is the intent: is there for example a commercial intent here? If the search would be “buy black aline dresses” that is easy, it will for example get a quality score of 10. Other searches like “”Buy black aline dresses with sequences” ail also be relevant, but less so giving it a Quality Score of 7,8 or 9.
As Bryan says in his keynote:
“Landing page relevance is not just about matching a keyword but about satisfying intent”
Google uses all the data they can get to understand both the searchers intent as the quality of the page, what is on it, but also how people (will) act on the page itself. That is where ‘big data’ will pay out for Google. They have 450 million new searches every day which will give them more information, so that data is growing by the minute.
And the more data Google gets the more it can play with it. For the sake of argument let’s look at Google as a casino.
If people go to a casino they usually have a pattern, they will do the penny machine first, then go to the dollar machine, then the nickel machine and back to the penny machine. If you are a gambler you know you will not win the big money this way. The big money you will win by sticking to one machine. Casino’s know this and they want you to move around to make sure you spend the most money. But that means they will have to keep you interested. If you wouldn’t win anything on the different machines you would bail out soon enough, so they make sure you win small amounts of money on different places, just to keep you going.
Potentially Google can do the same. By getting more and more data they will know where people will want to go and they can move around ads and results knowing people will click on the right ones to ‘keep them going’. Now Google will not be doing this in the ‘evil way’ which most people expect, but in a way which will give people the right answers.
Now this will cause a major problem for the marketer. The question after all is: if Google
gets so much new data and so much more understanding of the user every day, how can the marketer keep up?
The marketer will have a very hard time actually personalising the content, making it exactly fit for the intent of the search and thus Google (both paid and organic).
This means the marketer will have to start automating as well. Get different types of data and be ready for different types of intents. Pages that will adapt based on the search and how people come in. Pages that adapt to the searchers intent to satisfy the visitors intent.
Many marketers will have a hard time adapting to this because they still look at marketing ‘the old fashioned way’. When it comes to SEO its for example focussing mostly on links. But to keep up they will need to automate.
So will they be replaced by automation? Partly but there will still be plenty of things to do for the marketer, don’t worry about that. Bryan Eisenberg is not afraid to say that “Big data is going to replace the majority of us.”
There are already tools like Datapop who are acting on this, will marketers be able to adapt?
If you want to know more about this make sure you get the whitepaper from Bryan and Jeffrey Eisenberg on this here. In that they for example show how Amazon is working with the big data to get this working.