I had a first suspicion when I compared our previous CPC campaign, managed in house with the one who was done by agency. In previous campaign there was no traffic increase from organic traffic, while in the campaign managed by agency there was a large and correlated increase of traffic from both - CPC and organic traffic. Look at the image.

I have looked into various things in Analytics, but the strangest thing was appearance of CPC searches with plus sign, like +very +best +product. Now I doubt that there are lots of people doing search like that. Also, here was significant increase of average duration and pages per visit.

We did some other marketing activities at that moment, but it looks like a very unlikely coincidence.

Please advice how can we bust them, because my boss is about to sign up a contract to do SEO with them...


I think you have made a big mistake.

What you are most likely seeing is finally the result of a cleverly run AdWords campaign. The suspicious words +very +best +product that you dislike are almost definitely the keywords used in the AdGroup.

They are known as Broad Match Modifiers and bascially they dictate that all the words marked with + must be present to activate your ad. This negates a lot of unwanted traffic and will help to only drive traffic that is interested in your product. These visitors are way more likely to spend more time on site and convert.

Increased organic is hard to manipulate too. In your case it likely means that your product earned a lot of awareness via the AdWords campaign and it resulted in natural traffic.


For a beginning let me say that if they're good you won't be able to track them regardless how hard you try.

How I'd approach the problem is by installing an independent statistics system on the website that allows IP tracking.

Then look for patterns - if there's a lot of traffic from certain IPs: try to match them against agency.

Keep in mind that agencies usually hire a lot of different services. So your best bet would net to use whois and see if anything is registered either for an agency or any of it's employees. Keep in mind here that a single agency might operate as much as over 50 different IP addresses that usually will show as registered to the hosting company, not agency itself. One attempt here you might try to do is matching the whois entries with hosting provider of the agency website or any of it's services.

Another option would be to use geo-localization from Google Analytics and try to see if most of the traffic doesn't come from one or two cities. It might not work though if your website targets specifically into a single site.

Your last hope might be an attempt to find unusual patterns in desktop resolutions or browsers and compare that with average statistics for your country. Say: If 50% of your traffic comes from one specific browser running one specific resolution then something clearly isn't right. But in the end it's hardly a proof that they cannot dismiss.

As I said - it's really difficult problem to solve. I encountered it 3 times so far, and managed to prove guilt only once, and only because the service provider was from small city in the other edge of the country, so it was quite obvious that they generate traffic themselves while looking at the Google Analytics. In other cases I didn't have any proof that would be hard enough to make any demands or go to the court. Hope it'll help though!

  • Most of the things looks pretty random and/or inconclusive, although there is increase in use of one particular browser, which tops our statistics anyway.
    – dzh
    Feb 6 '13 at 12:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.