Generic domains as an Adwords campaign booster...

[Edwin]

There's a thread over in Domain News about a study I conducted, but since it's all about Adwords I thought I'd start a thread about it here...

Basically, I tested 3 URLs against each other, a totally generic URL, a quasi-generic URL, and a non-generic URL. All the other parameters of the ad campaigns were the same (ad headline, copy, max bid, landing page etc.) - only the domain names differed.

The result: using the generic domain name that was an exact match for the topic in question (electric bicycles) I was able to get up to twice the traffic as I was able to generate from the ad that used the non-generic domain, and it handily beat the quasi-generic too.

Here's a link to the full report, with charts, stats and graphs...
http://www.memorabledomains.co.uk/ppcanalysis.pdf

Incidentally, it's no longer possible to conduct this kind of test as Google has changed the rules so that you can only have one domain name per Ad Group - I just managed to scrape under the wire before they implemented the change, and collected quite a lot of data...

 

[foz]

Great stuff Edwin.

I do remember Adwords bringing in the rule that the displayed URL had to match the destination URL, as I assume advertisers were using domain names that they did not own or have control over. e.g. a flight booking branded website using www_cheapflights_co_uk in the ad.

 

[Edwin]

Just to clarify, I used only domain names that I had control over.

The topic of the test was electric bicycles.

The domains I used:-

electricbicyles/co/uk
yourbikes/co/uk
inahurry/co/uk

Google used to allow ads for any number of different domains within one ad group, so at the time (Feb 2009) I was able to run all 3 of the above directly against each other for an accurate comparison.

 

[lex007]

excellent post...cheers mate!

 

[miked]

Yes, excellent study. I've linked to it from my domains site... hopefully this type of evidence will convince more people to buy generics, and thus benefit us all!

Mike.

 

[stevebrowne]

great stuff, Edwin. I assume you will be giving this out at the Internet show as well?

steve

 

[Edwin]

Yes. I've asked my designer to polish up the document a bit so that it looks nicer printed out, and I'll probably run off a hundred copies of it or something. In any case, the salient figures and the main graph are reproduced in my general show brochure, so the full doc is really just to hand to SEM firms and the like.

I'm running out of time for preparation, but I'm hoping to have another couple of juicy handouts at hand for the show too...

 

[Darren]

Excellent post and well done on all the hard work promoting UK domains.

I look forward to viewing your stand at IW2009.

Thanks, Darren.

 

[julian]

nice but..

nice post Edwin but there is perhaps one flaw in your hypothesis...

in your examples you used

electricbicyles/co/uk
yourbikes/co/uk
inahurry/co/uk

I think this test is skewed to far in favour of the generic.

If I saw the same advert you used with the url:

cheapelecricbikes/co/uk - i'm probably guessing that I would probably be inclined to click that ad lots?

also if I was a user living in London sat at my desk and saw the same advert with the url: elecricbikeslondon/co/uk - i'd click that ad lots to.

hence the problem is these domains can be bought for reg fee - obviously electricbikes/co/uk is better but I often find that end users won't pay the prices and will go and reg some combo like the examples above - and infact if i had an electricbike shop in london londonelectricbikes/co/uk etc is probably better?

Your point overall is good but I wonder if the penny dropped with end users yet, and the if so the more savvy they get means they will just reg more savvy domains themselves..which would mean doing a monopoly bulk reg'ing job on generic combos?

 

[devolution]

Edwin - something I noticed was on your bar graph on the last page, it says that the Ads in positions 2 & 3 did better than in position 1, which is surprising?

 

[Edwin]

Edwin - something I noticed was on your bar graph on the last page, it says that the Ads in positions 2 & 3 did better than in position 1, which is surprising?

The graph is showing the percentage of clicks that came from that position as a percentage of all clicks received.

It's impossible to see how many impressions were served at each position (Google doesn't publish that info anywhere) so it's not a graph about CTR, but I thought that while it doesn't tell us that much in absolute terms, it's the relative comparison (between the 3 domains in the study) that is interesting.

 

[retired_member34]

It's an interesting study, and great to see someone taking the time and effort to document it.

However, I do think there is bias in the method used, particularly in your Group A. It is well known that using the entered keywords in the ad will raise the CTR. You can do this in the ad title, the ad body, and the display URL. You have used the keywords in the ad title and ad body across all the ads, but only in the one display URL, that of the generic.

There are other methods though. To get a clearer comparison I would have used the alternatives yourbikes/co/uk/electricbicycles and inahurry/co/uk/electricbicycles. (Similarly in Group B I would have used electricbicycles.co.uk/electricbike, yourbikes/co/uk/electricbike, inahurry/co/uk/electricbike)

I still think the generic domain would have performed the best, but the difference would not have been so great as recorded.

I have long used this technique with success. Another tip is that if you are bidding on TM keywords, Google stops you using the keywords in the ad title and body, but does not stop you using it in the display URL.

 

[Edwin]

Fair points. Had I had time for a longer study, I'd probably have gone in all kinds of directions, as suggested - but Google effectively pulled the curtain down on proceedings...

 

[doodlebug]

Good timing with your test, yeah a generic will almost always do better but we all still need to convince end users.

As for a London shop selling electric bikes electricbikesLondon would be good but this shop as a website and potential to sell to a worldwide customer base, so this only appeals to a place with a population of approx 8 million and not for eg, the rest of the UK with an approx population of 55 million
Overall electricbikes/co/uk would be a lot better

 

[ddwebguru]

Generic domain may not be a decisive factor in Google Adword but it plays a crucial role in Google web search.

 

[bensd]

Generic domain may not be a decisive factor in Google Adword but it plays a crucial role in Google web search.

Edwin's study shows that the generic domain was a decisive factor as that performed better...

 

[Edwin]

Edwin's study shows that the generic domain was a decisive factor as that performed better...

Yeah, that was kind of the whole point!

 

[bensd]

There have been so many of these these types of posts on the forum lately, not really spam but i can't help thinking they are only trying to index the links in their sigs.

 

[julian]

I think your wrong because first of all to dumb down 'only 8 million potential cutomers' as if everyone man and his dog has 8million, and secondly you probably would'nt be sending electric bikes through the post to Beijing so you don't need a global market for a product such as these and although I like generics in my experience people will buy and use what they're given - provided it has the product they want ..

Good timing with your test, yeah a generic will almost always do better but we all still need to convince end users.

As for a London shop selling electric bikes electricbikesLondon would be good but this shop as a website and potential to sell to a worldwide customer base, so this only appeals to a place with a population of approx 8 million and not for eg, the rest of the UK with an approx population of 55 million
Overall electricbikes/co/uk would be a lot better