I just got off the phone with my friend.
Here’s what he’s doing:
1) Launching 5 new CPA offers per week in various categories by direct linking.
2) Each offer contains 200 keywords, including domain names as keywords
3) Bid price = Expected Conversion Rate x Payout
The basic strategy is to find winning campaigns fast using direct linking, then build out a landing page and full affiliate site.
He hasn’t had much success with this method, so he asked my advice about why it’s not working so well.
I was happy to answer this question
First, I asked him what his budget was for his latest campaign he launched. He told me $50. I asked him what his CPC was, he told me $2.93. So he basically got 17 clicks before he turned off his campaign, since none of those clicks converted.
Before I go on.. Do you guys see what’s immediately wrong with what has happened? This is a good exercise to think about for a second before reading on…
Basically, he’s bidding way too high. If he bids lower, he can drop his position down to 8 or so on the page and get twice as many clicks for the same budget.
Positions lower on the page also convert better.
I wanted to drill down further, so I asked him to look at the keywords which received clicks. He told me there were just a handful, maybe 4 that accounted for all the clicks.
This is another BIG problem.
Because those 4 keywords ate up all his budget, the other 196 keywords didn’t even have a chance to get clicked!
Which brings me to my second point– why is he testing only 200 keywords?
He was told that he should crank out these campaigns fast, and you want to bid on the top 200 keywords- if they don’t convert well then move on.
I think this is bad advice.
You may ask me, why would testing more keywords be beneficial?
I have a few logical reasons. Now that you have 1000 instead of 200 keywords in your campaign, you can bid much less and still get the same amount of traffic.
My first piece of advice was to lower his bids A LOT so that he can get double the amount of traffic for the same budget. My second piece of advice is to increase his keyword size to 1000, so he can again lower his bid price and receive the same number of targeted clicks for his budget.
As long as his keywords are highly relevant, I see no negative impact of adding more keywords to a campaign. Granted, it does take a bit more time to set up, but the trade off is well worth it.
Now, my third point.. (Hopefully I’m not boring you guys, and you are still following me here!)…
Does your test have statistically valid results?
The fact is that his budget was so small that he could not produce statistically relevant results, even on the 4 keywords that received all the traffic.
But nonetheless, let’s assume he did establish statistical significance with a $50 budget on a few keywords that got most of the clicks (assuming his bid price was WAY lower), can you then extrapolate that to mean the offer is or is not converting? Doubtful.
Finally, my fourth point…
Again, this has to do with “what are you REALLY testing here?”.. He’s basically testing a set of keywords from a particular traffic source going to a particular landing page. So even if he followed my advice and reduced his bids, and increased his keyword list to 1000, he STILL does not know if this offer converts.
Why? Well, he’s only testing the merchant’s landing page. Yes, you could find the occasional super-star of an offer that converts with the merchants landing page, but that’s the exception and not the rule. So he has to understand that he’s not really testing if the offer is even profitable, but if it’s profitable with one PARTICULAR landing page.
So now that I’ve gone on a long rant, I want to clarify. I love the idea of becoming more familiar with Adwords. In fact, I highly recommend the strategy of testing a ton of offers using direct linking, because you will learn how to drive traffic effectively and affordably, which is a critical component of success in affiliate marketing.
In short, all I’m saying is:
1) You also need to know that it’s better to bid lower and see the traffic level and increment your bid up, rather then blowing your budget on a few keywords when there’s a huge amount of search volume.
2) Bidding on a BROAD set of highly relevant keywords is always a good strategy. Artificially limiting that with a hard rule that says “only 200 keywords” is a bad move.
3) Understanding the importance of ‘statistical significance’ is very important to affiliate marketers trying to conduct accurate tests.
4) You need to be aware of what you are IN FACT testing (scientifically speaking). (Ie. What you are really testing is: will this traffic source convert with this particular landing page with these particular set of keywords.. and not “Is this offer profitable?”)
Peace be with you!