I just watched a fascinating video by Rand Fishkin on predicting keyword volume. It’s geared to his neck of the woods (SEO) but was important and insightful enough that I wanted to shed some additional thoughts on predictive keyword research for PPC.

What is predictive keyword research?

To me, it means getting ahead of a term or topic before it becomes popular. Think about the term “attribution modeling”. Before a few years ago this didn’t show up in search at all. Today it is all the rage and those who built the first landing pages, ads and bid on the keywords we ale to ride that keyword to some lead generation success until it became more crowded (currently 8 ads for that search query and the average CPC is projected at around $3).

Google Insights for Attribution Modeling:

Predictive PPC Keyword Research

Why bother with PPC keywords that won’t get any traffic today?

Well, you probably shouldn’t if all you are worried about is today. But if you are in it for the long haul then these keywords are where your account will find it’s future growth. Getting them in your account with the proper ad and landing page will help you build history, you will have a higher CTR once people do start searching for it and this a higher quality score and you will start getting the conversions as soon as people start searching for these terms and thus will aggregate a higher number of total conversions versus being the last to the party.

How do you conduct predictive keyword research?

Rand recommends a few methods. These include staying up to date on news alerts, social sources and checking in with Q&A sites to see if anyone is asking follow up questions about the news and social media topics. These all make sense to me and are a good place to start but certainly not the only place.

Here are a three additional ways to predict if a keyword that isn’t currently popular will gain some traction.

1. Look at conference topic titles. Do any stand out as having some potential? If you saw a presentation titled, “What is Local Search Marketing” by a prominent speaker and you had a local search marketing service you may consider getting ahead of that curve.

2. Take sales calls. Then look for patterns in what they are selling. If a company is investing a lot of dollars in outbound telemarketers they may be 1, creating interest in a unique service that you could capitalize on or 2, they have already identified a trend and are trying to jump on it quickly too.

3. Movies and TV. This one is the best. Mostly because it involves you killing some brain cells. But also because movie and TV terms and products spread fast. Think of Luwak coffee. Did anyone know what that was before the movie bucket list? (By the way, it is delicious.)

For eCommerce this can apply to potentially new products to start stocking, to keyword qualifiers that you should be bidding on or negative keywords you should add because the trend isn’t what you sell but could get spill over traffic (think someone who sells Olympic watches during Olympic season).

Lead generation companies can use predictive keyword research to find cheap clicks and take advantage of big spikes in traffic when everyone else will be rushing to get in on the action after it is too late.

In addition to the ideas I have mentioned, I am looking for additional tools to use for predictive keyword research, any of you have suggestions? Leave them in the comments below, thank you!