Data: long tail keywords are coming to an end

Category: The Internet for Human Beings | 4 min read

A few months back I was at a Google hosted Fintech event. At the event Google dropped a pretty cool statistic:

Over the past two years, searches for local places without the qualifier "near me" have grown 150%.

That was quite the surprise for a lot of us marketers at the event. They then went on to their usual talk about micro-moments.

This  quote got me thinking about how people search in general. I thought  about my paid search campaigns and how I had noticed that my long tail  keywords had been dropping in traffic while the generic terms had grown  exponentially.

Being a data focused guy, I decided to explore this idea a bit: "do people need to use long tail less these days?"

My  theory was that Google had gotten so good at search that people simply  trust that Google will give them the right answer, even without  explicitly saying what they are looking for.

First, some data

I  work in the insurance and finance industry so lets use an example from  there. The terms "cheap auto insurance", "auto insurance quotes" and  "compare auto insurance" are far from long tail but they are a lot  longer than "auto insurance" and also give far more insight into the  searchers intent.

The issue is this chart:

The  data is specific to Canada and taken from Google Insights. As you can  see, both terms have been dropping pretty heavily. Here is what just "auto insurance" looks like:

I  can tell you from experience that this is not a simple seasonality  chart. Now for "cheap auto insurance" again, something with more intent  information.

I  think you're probably getting the point by now. People just aren't  using the keywords that even in recent years SEO companies would be  telling you to use.

The classic lesson they teach in SEO, and one  I still teach to this day to university students is the importance of  intent. What people type into Google can help you make sure you are  getting those with buy intent or at least know where in the buying cycle  a customer is.

When customers now just use "auto insurance" it doesn't really tell you anything.

This all matches what I am seeing in my paid search campaigns as well.

Okay, but why?

I  think my theory above is correct. First of all, Google is getting  better at figuring out the intent for you. With "near me" the intent is  clearly looking for something nearby, Google has gotten to the point  that if you type in Italian restaurant, it will auto use your location  to find some.

It started with stuff like nearby and the power of Google's LSI algorithm. These days though they have moved on and are using machine learning and  AI to guess the intent of the searcher. This can be based on data you  don't even realize. Knitters for example. I was shown by a Google  employee an example where someone they knew was a knitter would have a  counting app suggested to them. Something a normal person who wasn't a  knitter would not be able to intuit. Eventually it was realized this is  because those who knit need something to keep track of the rows and  knits they were making and the app was super helpful for that.

People  are also doing far more searches on mobile devices. While mobile  devices are pretty easy to type in these days, they are still more  challenging to type in than an old fashioned keyboard. People want to  type less because it's annoying and since Google gets it right most of  the time anyway, they get the excuse to not type as much.

What do I do?

You have to change your overall game plan.

In  terms of paid search the good news is that with automated bidding you  don't need as much intent information. Google figures that out for you,  it's more important to get better data to Google so it can do that for  you.

SEO? Well, that's a tougher nut to crack. In the past long  tail was the place you could build a whole business. Combine the lack of  long tail with Rand's data here about no click searches and you have a serious problem for entire business models.

I would suggest taking a look at my other article that talks about clicks in general here.

The focus of digital marketing needs to switch to be more about brand and being top of mind. If you look at the data here, it shows that people are FAR more willing to click a search result if they already know the brand.

That  can get expensive though. It becomes a question of how do you target  your top of funnel brand advertising in a way that is cost effective? I  don't think anyone has answered that just yet. Combine that with recent  failures by large companies such as Google at proper attribution  tracking and the ad fraud issues that are rampant in the programmatic  landscape and you have brand new challenges that need to be taken on.

Other Articles