Booking.net.au – Experiment Part 2

Posted by: Robert Austin on Mar 12, 2009 | No Comments

So the booking.net.au experiment has been running for about 1 week now.

Heres some updates.

Initially I had a lot of problems with ad’s being disapproved because the display URL did not match the Destination URL. After trying many combinations to get this to work only 1 method passed approval and allowed me to keep a display URL with the /region visible. For example Booking.net.au/Brisbane/

  • I had to disable 301 redirects through .htaccess
  • I had to create physical .html pages for each region. See www.booking.net.au/brisbane/
  • I then entered the affiliate link at the keyword level “destination URL”

Using the destination URL at the keyword level is a popular trick but is kinda crap because you lose the flexibility of changing your campaign on the fly through the .htaccess.

Anyway, first problem solvered.

The 2nd issue I’ve run into is low bids. I mentioned in the first post my maximum bid would be $0.50. Pretty much all generic keywords like “brisbane accommodation” etc won’t even trigger with a bid this low. So I’m stuck with pretty low impressions overall.

Why not increase the bid?
If you can make enough money from commissions with a PPC bid higher then 0.50 on a setup like this then your 1 of a lucky few.

Using my magic forumla for profitability you will see we have a very hard margin. This formula is actually based on a lot of real data from similiar previous campaigns and is a great rule of thumb. It relates to generic travel keywords and niche keywords. It does not include brand. Also it is landing customers on a website which is a brand they have never heard of, and has no level of  trust established.

100 clicks = 1 conversion
1 conversion = insert average commission

So based on my max CPC of $0.50
100 clicks * $0.50 = $50
My average commission will need to be higher then $50 to make a profit. And let me tell you that’s a stretch.

Here is a screenshot of today’s search volume and clicks from Google Adwords.

In total I have spent $100, generated 200 clicks and made 2 commissions. This is the hard margin formula to a tee.

Total commissions worth $79. average commission $39.50

I have produced an average commission of $80 over 2-3 month periods with a similar setup so I’m not writing this campaign off yet.

What is my prediction for the campaigns profit by march 31st? Loss of $150.

Booking.net.au – Travel Affiliate PPC Experiment

Posted by: Robert Austin on Mar 5, 2009 | No Comments

The idea: create a scalable PPC campaign promoting an affiliate travel site using only redirects.

I’ve allocated a budget of $1000 for March and the goal is to generate a profit.

There are  no physical pages except for the homepage. It just redirects from Google Adwords into the affiliate booking engine with a custom theme.

This is not a new idea, many have tried it, dreaming of the easy money and 2 years ago something like this was an awesome money spinner.

But rising click costs ( because of increased competition ) and increasingly strict Google guidelines & affiliate guidelines made it unprofitable ( well mainly the click costs made it unprofitable, the latter just made it more difficult ). So years later, I’ve returned to see if holy grail still exists.

Register and host booking.net.au / create logo
It was important to register a short domain name because I want to use regional sub folders in the Google Ad text.
For example Booking.net.au/Brisbane
It is my opinion that using regional sub directories in the ad text will slightly increase consumer confidence and quality score.

Setup .htaccess file for redirection
I used a .htaccess file to control the affiliate redirects for maximum future flexibility. Also it’s pretty easy to use excel to create all the redirect combinations and then copy and paste into the .htaccess file. If you have a spreedsheet provided by the affiliate program its even faster.

Setup Google Adwords campaign
I’ve opted to start by pushing generic accommodation keywords grouped by region. A generic region would be “gold coast” while a sub region might be “broadbeach”
The core accommodation keywords are “accommodation”, “hotels”, “resorts” & “apartments”

Generally the volume of impressions also goes in that order, “accommodation” recieving the highest and “apartments” the lowest. Sometimes this changes based on region. For example CBD’s and major cities normally have lower impressions for “resorts” and “apartments”. On the flip side  “hotels” will have a higher impressions, sometimes out pacing “accommodation”

For those of you interested you can take a peek at my campaign structure exported directly from the adwords editor.

Why target generic accommodation keywords?

Mainly because I’m interested in seeing if the margin is there. It’s also natural starting point. If costs are too high, or the margin is not there, it is a more logical process to sub divide and chase the long tail from this point.

My predictions are:

  • Generic accommodation keywords are going to be expensive. Probably to expensive to generate a decent profit margin.
  • Generic accommodation keywords have the advantage of landing the user on a page with more choices. Which in theory should increase the change of a conversion.

Will the increased number of conversions counter balance the higher click costs. I doubt it but that’s the whole point, to experiment.

Bids & Targeting

So far the campaign targets most of Queenslands most popular destinations. I’ve set the max bid to $0.30 so it’s quite likely I wont even get 1 impression after I go live.

Upload & Go Live
When your working with large campaigns it’s critical to segment campaigns and adgroups and not overload one particular campaign with to much data. If you try and upload it all in one hit the adwords editor will crap out and you will be stuck with errors that can be difficult to get rid of. Just say you upload 80% of a massive campaign and it stuffs up, if you then go and try and delete it so its smaller, it will stuff up again, because its trying to delete to many ad groups and keywords, get it?

So segment carefully and upload in blocks. To be honest, this is actually a better approach as well because your going to take more care with your campaign.