Up and down the country in marketing departments far and wide, people are summarily passing judgement on keywords and executing them from campaigns because they don't cost in, do the numbers, make the results, and get the returns.
But not many marketers have the time to ask why. The naive math of it all says, bid it down or cut it off, and whilst that may stave off a problem in the short term, it's often not really the keyword's fault. It's the parasitic activity within it, that you can't see that lead to good keywords being misjudged.
What I'm going to cover in the coming short series of articles is where to find parasitic search, that's bleeding your ROI in your paid search activity. I'll show you how to weed it out to improve your ROI, without damaging the performing elements of the campaign.
Unless you're exact matching everything, which is unlikely and not a little bit dangerous, the broad and phrase matched keywords you're bidding on are not actually what you are bidding on. They're wildcards. They are gateways to traffic that is both identical, similar and sometimes a bit of a way off the context that was your intention to capture.
And because it varies in its context, it differs greatly in quality.
In fact in a few accounts we recently optimized, the actual traffic that could be determined as being of the high quality that was expected, was less than 60%. The other 40% was extended query clutter, ie other searches that fell into the realm of a keyword they were targeting using a broad or phrase match, and triggered their ads. It's this extraneous activity that is invisibly costing you money.
If there's one thing I've learnt from nine years in search marketing, it's never under-estimate the capacity of people that will put the words Britney Spears onto everything. Not to mention the eternal optimists that want everything from an ipod to a family vacation to a Maserati, for 'free', the spongers looking for pics to hotlink and what not and so on.
And that's just the obvious stuff. The range of extended queries you will capture out of broad and phrase matching can be vast and granular, and there are ways it hurts your campaign.
Even for the good phrases, this clutter comes at a higher price than it needs to. When you broad or phrase match and set a price per click, you are also setting that price for any extended query traffic caught by it, when in most cases that traffic could be purchased at a fraction of the price if that query were to be set individually as its on keyword and bid on.
For example let's say we're a DVD retailer, bidding on 'Star Wars dvd' as a broad. We get plenty of traffic for the term, but we also get a lot of traffic for 'Star Wars dvd boxed set collector's edition', 'Star wars widescreen dvd' and 'Star Wars dvd boxed set'. It's all good traffic, and probably quite relevant if we carry those lines. The problem is we've over paid massively for those clicks, because 'Star Wars dvd' is a popular term and we're really running at a high cost. Maybe we're even running it as a loss leader because it's just out and we want to buy a bit of market share from everyone else who's out there doing the same.
In many cases this traffic could be acquired as much as 80% cheaper. This brings us on to the second reason.
A lot of those extended queries that you're over paying broad match prices for, may be irrelevant and useless to you. Continuing with our 'Star Wars' example, where we bid on 'Star Wars dvd', we also picked up 'Star Wars dvd rip' and 'Star Wars dvd torrent'. These are not terms we're interested in and the offending words need to be added to our negative list pronto (if we haven't already got them in there) to cut down on the re-occurrence of our ad showing for these terms and wasting our click budget.
Now this is going on right now in your campaign, and you can't build every word into your negative keyword list. This is something that needs continual refinement and effort to stay on top of.
We recently optimised an account for a nationwide automotive dealership for their first time. We found a vast amount of traffic being pulled in under towns and they didn't serve within a 20 mile radius, searches for all sorts of parts and spares, wreckers yards, repos, insurance write offs and all manner of things that most people just wouldn't have thought would have materialised. Over the course of the last year or two, these searches had generated millions of impressions and thousands of clicks, and at a cost of tens of thousands, with suffice to say very little to show in terms of sales. Not to mention the havoc it plays with your quality score and the cost implications of that (but that's a whole other article).
So where do you go to get this vital data on the extended queries that are triggering your ads? Well there are two key places to look. Your web analytics system and what I'm going to discuss in this article, which is a little known corner of the adwords interface called the search query performance report.
You'll find it in your adwords account under the reports tab,
and you'll need to select the search query performance report and the date options over which you want to search.
Once you've done that you'll end up with a report like this. Here's one we took for an automotive client of ours (I've cleaned this up a fair bit to illustrate the point)
What this shows me in the various columns, are the extraneous keywords (Col 3) that I'm receiving ad impressions and clicks for (Col 5-7), because of keywords I'm broad matching (Col 5) in my various Campaigns and Ad Groups (Col 1 & 2).
So now if I take a portion of that data and look at it closely, I can see some things that should raise alarm bells, like the keywords that are little to do with our clients business of selling new and used cars.
Collector words like classic, irrelevant terms like Golf cars, damaged cars, and places like Glasgow where they don't operate, are costing us clicks and money, and the low CTR we get for those terms also pulls down our adgroups quality score and could inflate our click costs overall.
These offending terms and words need to be added to our negative lists at the various ad group and keyword levels.
If you've got a spare 10 minutes, take a quick look at this in your accounts. It's the kind of activity that will start saving you money the minute you do it, and is one of the largest sources of parasitic search we come across.
You'll need to do this task over a large historic back period first and then keep on top of it weekly, although few people actually bother to do this, as there is a large amount of work generated by this task, as you'll see when you have a look at your report.
The best way we found to deal with this consistently and regularly was to develop an automated system that pulls this data in regularly and scores each extraneous query by looking up a taxonomy to determine if it should be negative matched or created as a new keyword or even adgroup.
If you can clean this body of data out initially and then regularly after that, you'll start to see some immediate and noticeable improvements in your paid search performance.
There are other ways to deal with this too, using your analytics system. So stay tuned and we'll go through how to do that in our next newsletter.