Ask Me Anything ClickBank Algorithm

[Graybeard]

SAMPLE>>>
{
"Category": [
{
"Category": [
{
"Name": "Art",
"Site": [
{
"ActivateDate": "2010-12-12",
"AverageEarningsPerSale": "30.5468",
"Commission": "75",
"Description": {},
"Gravity": "3.93434",
"HasRecurringProducts": "false",
"Id": "CHRISIA2",
"InitialEarningsPerSale": "30.5468",
"PercentPerRebill": "1.0",
"PercentPerSale": "75.0",
"PopularityRank": "1",
"Referred": "98.0",
"Title": {},
"TotalRebillAmt": "0.0"
},

Vader@DeathStar10:~/affilate_mainstream/clickbank/marketplace_feed_v2.xml$ grep -c '<Id>' marketplace_feed_v2.xml
9512 *offers* <<<

=================

Yesterday I started to parse out the ClickBank XML Marketplace Feed.
First I took that XML and encoded it into a JSON format -- IMO JSON is easier to work with.
I designed an initial algorithm and I have been testing the results last night and a bit this morning. Don't even bother asking what the algorithms are -- just like elGoog >>>I will just speak in generalities.

Here is the problem: Somehow I need to validate these algorithms.

So here is the deal: I am indexing these by the site ID. That should correlate to the name of the offer or the offer's URL if you are checking offers at ClickBank search. If you would like my algos ranking -- post the offer and ill try to find it.

• You MUST furnish the Category -- that is how the feed is indexed: Category-> Site -> Id

Remember, these algorithms are for research purposes only -- your success is not guaranteed AS-IS only rankings -- this is very Alpha now.

That said, upon visual inspection of a dozen or so offers, to me, subjectively and untested with ads or other marketing; 30% of the offers I found seem plausible and worth more due diligence investigation before developing marketing ideas. My inital reasoning is to narrow down the possibilities to the best offers based on hard data I have -- and then visually investigate -- every horse is not a winner and some are REAL LAME
How To Lose Man Boobs Naturally ranks well maybe there is a market

The best thing is that I have found what *may be* some sleeper offers with this algo (already) that sales may be infrequent for but the payouts are over a few hundred dollars with rebills. These offers are intended as long term earners in this particular algorithm. *Evergreen* as well as flash-in-the-pan offers that may be good for native ads >(controversial )

I am also going to design a low sale price and hopefully high volume algorithm later -- more geared toward mobile sales -- including free signup offers too: CPI, Surveys, app installs.

I chose ClickBank for this *experiment* for reason of easy (and known) access to a data feed. Other suggestions would be welcomed also as the same basic logic would apply

So, post offers at ClickBank to this thread, and time allowing, I'll post the current algorithm's ranking for free in exchange for some feedback -- limited offer yadda, yadda, bla, bla, bla ...

Also, your ideas for algorithm factors are welcomed and if you come up with a good idea you will get free access of some kind, when and if, I go production -- maybe with a SaaS based on a larger number of feeds. Or, compile my own network feed?

 

[Graybeard]

#12 on my list (so far)
This guy is good

He just needs a gold chain like a rapper wears

 

[Tarun Reddy]

Your algo will give the ranking, but any other data it can pull?
For example, here in this page ClickBank | Databank | Internet Marketing | ClickBank Tips they show the top conversting offers by offers, niche, pricing, countries, device etc.

 

[T J Tutor]

Crestani is very good at earning, but everything he does is black-hat. He is essentially teaching people to learn how to scam by making fake reviews on products and services that they have never tried.

 

[Graybeard]

When they publish that whole dataset (conversions) I will look at the raw data myself. They probably consider that proprietary -- understood.

That link is a PDF. I have a look but real numbers matter to me. I looked -- basically useless to what I am trying to do.

 

[Graybeard]

Crestani is very good at earning, but everything he does is black-hat. He is essentially teaching people to learn how to scam by making fake reviews on products and services that they have never tried.

I knew he was full of it -- look at the set, LMAO ... Empty offices after hours HAHA
I liked the good hype though
"making fake reviews on products and services that they have never tried."
*That's an accurate description of most review sites*

 

[Graybeard]

I contacted ClickBank and asked if they would make available to me more detail information about sales volume, buying countries and countries that had success making purchases on these offers.

They said no however I could contact the sellers for this sort of info if I wanted. IMHO that would be an unverifiable source of information so in designing a useful algorithm is would not be accurate. We are talking algorithm here not a *polygraph*

What I am finding: some old offers that may have been valid but didn't work out mainly for reasons of marketing failure -- the ideas make sense. I think some of these ideas could be 'remarketed' more effectively -- I see value in that down the road