Business Thread... Am I the only stats freak in the industry?

[Shap]

Hi. This is something that has always baffled me. There is so much information at our finger tips and yet in most cases the only thing that gets reported are Sales and Hits. In the last 5 years there has been lilttle development in the information gathered and reported. This goes from Processors to affiliate programs. Whether you are a tgp or a paysite owner. The truth is their is so much information we are all ignoring and as a result it's costing us all dollars. Using that information would allow us all to really start optimizing our earnings.

I guess my question is, does this bother only me and my staff? Am I wrong or are we all letting alot of $$$$ slip away?

[p1mpdogg]

I bet Jim is more of a stats freak than you.. but thats not a bad thing

[marketsmart]

i am a total stats whore. i crunch numbers six ways to sunday.. sometimes i go to far with stats and make changes before i give things a chance...

[Drake]

There is always room for improvement, you're right

[TheSenator]

Stats are valuable.

[Snake Doctor]

You know which company is a real stats freak?
Wal-Mart.

They have so much data at their fingertips it's scary, and they use all of it to generate more sales and cut costs at every stop on the distribution chain.

We could learn something from that

[corvette]

i agree

i.e. form hits/submissions/approvals are stats that (in my experience) do not get looked at enough

many clients just look at the # of sales

[bigdog]

Truthfully if you were really a stats whore wouldn't you be using nats or mpa3 by now

[Shap]

Originally Posted by Lenny2

You know which company is a real stats freak?
Wal-Mart.

They have so much data at their fingertips it's scary, and they use all of it to generate more sales and cut costs at every stop on the distribution chain.

We could learn something from that

Damn straight. I'm sure the stats google has would make alot of people's head spin.

[Shap]

Originally Posted by bigdog

Truthfully if you were really a stats whore wouldn't you be using nats or mpa3 by now

These things take time

[ElvisManson]

Originally Posted by shap

Damn straight. I'm sure the stats google has would make alot of people's head spin.

if you are looking for more stats check out these guys.

They have some decent products.

http://www.omniture.com/products/web_analytics

[CDSmith]

Anyone remember "The Statistician"?

"thestatistician.com" now redirects to a forum on the pond, but there looks to be a ton of statistical information compiled there.

[RogerV]

Back in the day we use to track everyhting from how many pages deep the surfer goes to what they mouse over and click on with a Java applet. In the end it was so random and the changes we made did not improve the conversions. Sometimes to much info is a bad thing. most of the time conversions go bad because of bank declines and its getting worse everyday. not much you can do about it.

If you do find the secret formula it doesnt mean it will do well because webmasters might not find it apealling. most webmasters now and days dont care how things convert they care if they look kewl or not on there free TGP's


I could go on and on been crunching numbers for 10 years.

[Trixxxia]

I'm a numbers freak too - I'm an excel freak too LMAO

[Shap]

Originally Posted by TopBucksTrixxxia

I'm a numbers freak too - I'm an excel freak too LMAO

I'm buried in excel at the moment. Have been for the past 2 days

[Shap]

Originally Posted by RogerV

Back in the day we use to track everyhting from how many pages deep the surfer goes to what they mouse over and click on with a Java applet. In the end it was so random and the changes we made did not improve the conversions. Sometimes to much info is a bad thing. most of the time conversions go bad because of bank declines and its getting worse everyday. not much you can do about it.

If you do find the secret formula it doesnt mean it will do well because webmasters might not find it apealling. most webmasters now and days dont care how things convert they care if they look kewl or not on there free TGP's


I could go on and on been crunching numbers for 10 years.

I understand. But let's get even more basic.

How many sponsors have the exact data of how many hits each hosted gallery gets?
How many have the click thru data.
How many clicks from the gallery hit the tour? And the Click Thru Ratio on every single hosted gallery?
How about sales. Hosted Galleries with the best signup ratio? Hosted Galleries with the most signups vs hits to the gallery? Or how about the galleries that received the least hits that converted best that should be pushed even harder?
Or something as simple as a list of their top 250 hosted galleries in order by sales or conversion ratio.

Having this information ready at our fingertips is absolute money in the bank. My guess is 90+% of sponsors do not have that data. I know we have some of that information and only because I have people working full time digging up the data for me. It should be all automated and WITHOUT causing server problems.

[GTS Mark]

Stats are great but I think some people spend too much time tweaking the hell out of their sites to get a 1-5% increase in sales and could spend that time creating more sites/programs and make a huge increase in their overall business.

If you want to take a look at someone that has spent their time creating more and more sites and less time perfecting every little thing on one site is Bike from Tiava. I believe there is a thread on here somewhere where he just purchased a Ferrari hehe ;)

DH

[Dollarmansteve]

Originally Posted by shap

Having this information ready at our fingertips is absolute money in the bank. My guess is 90+% of sponsors do not have that data. I know we have some of that information and only because I have people working full time digging up the data for me. It should be all automated and WITHOUT causing server problems.

Sponsor programs that run custom backends have lots of numbers to play with

[Shap]

Originally Posted by DrinkingHARDEST

Stats are great but I think some people spend too much time tweaking the hell out of their sites to get a 1-5% increase in sales and could spend that time creating more sites/programs and make a huge increase in their overall business.

If you want to take a look at someone that has spent their time creating more and more sites and less time perfecting every little thing on one site is Bike from Tiava. I believe there is a thread on here somewhere where he just purchased a Ferrari hehe ;)

DH

You only say that because that makes you feel like the leafs have a chance at winning a cup some day when statistically it's impossible

Seriously though, you bring up a good point. However I think milking the % here and there and optimizing is very important. Think of the money spent in designing galleries, in hosting the galleries and yet how many companies actually know which galleries are working? Our industry has a history of continually throwing traffic at things and hoping they work. The truth is if you show an affiliate bad sales they may move on. Running an optimized business can help you retain members and affiliates and make more money in the long run.

[Shap]

Originally Posted by Dollarmansteve

Sponsor programs that run custom backends have lots of numbers to play with

I figured that

How much of a problem does tracking more information cause with the servers?

[NoWhErE]

I'm 50/50 with you guys, of course stats are great, but I've known people to go overboard looking for that 1-2% increase in sales.

You need to find that fineline that seperates optimization and perfection.

Optimize a formula and use it like crazy. And of course, keep looking at your other formulas from time to time to see how you can improve on them

[Shap]

Originally Posted by NoWhErE

I'm 50/50 with you guys, of course stats are great, but I've known people to go overboard looking for that 1-2% increase in sales.

You need to find that fineline that seperates optimization and perfection.

Optimize a formula and use it like crazy. And of course, keep looking at your other formulas from time to time to see how you can improve on them

See I don't think it is 1-2%. I think with really detailed information the increase would be 7 to 12%. That is huge. You show your affiliates that type of increase and they'll increase traffic. And that 7 to 12% becomes a 15 to 20% in the long run. That's alot of $$$$$

[Pornwolf]

You are right Shap. Those kinds of stats are missing for the affilliate, but if the program owner doesnt have them something is wrong.

[Yngwie]

Originally Posted by shap

I figured that

How much of a problem does tracking more information cause with the servers?

well, I say that if it would cause server problems with more tracking than just have a server strictly for the stats shit. Would be nice to have more info available to us.

[JOELK]

Paying attention to stats is the only way to really be efficient. I work to make money and the most money possible. This is what i think mainstream companies do that we can all learn from

[WiredGuy]

Oh god I'd love to get more stats data. Things like number of CC submissions versus number of sales (figuring out what percentage of declines occured, CECash shows this). Things like postback's on keywords that resulted in sales, geographical sales data, etc. etc. Problem is, too much data can overhelm people as well.
WG

[BlingDaddy]

Originally Posted by marketsmart

i am a total stats whore. i crunch numbers six ways to sunday.. sometimes i go to far with stats and make changes before i give things a chance...

That describes me.

[BlackCrayon]

Originally Posted by NoWhErE

I'm 50/50 with you guys, of course stats are great, but I've known people to go overboard looking for that 1-2% increase in sales.

You need to find that fineline that seperates optimization and perfection.

Optimize a formula and use it like crazy. And of course, keep looking at your other formulas from time to time to see how you can improve on them

not sure why people would be concerned about 1-2% increase but you never know what tweaking could bring a 10 or even 20-30% increase unless you try.

i love checking stats and agree we could use more info, most specifically sponsors need to cut out all this second page, join page only counting and need to start showing approves/declines for starters. also sponsors who don't show the referring urls need to get with the times.

[ServerGenius]

Originally Posted by shap

I figured that

How much of a problem does tracking more information cause with the servers?

none if you do it correctly and very simple. Store logfiles on a mounted disk
on a seperate server. Use that server for nothing else than logfile crunching.
The same can be done with dbstats of course.

For updating stats in Mysql tables I suggest to make use of delayed inserts.
That's a LOT more efficient than updating every request and reduces the load
on your mysql server significantly

[davethetruth]

Great thread!

I was thinking about starting one just like it the other day

I used to be a stats freak, but since we put out a new version of our software, a lot of the program stats I used to have haven't been implemented yet and I'm pretty much going through withdrawls.

In my personal affiliate account for BRS I've got about 1000 daily link stats based on streams of traffic, but that's about it

[MetaMan]

we have a custom system being programmed into our tgp that sorts listing by convertions, click, sponsor type, processor, niche.

and updates changes across the entire network.

[Dagwolf]

I love my spreadsheets almost as much as I love gaming. :D I'm a compulsive stats-checker as well but I've learned to limit myself to once a day.

[Shap]

By the way, if there are any stats freaks or programmers out there with existing products or products in production drop me a line. I'm not interested in an affiliate program. But plug ins or things to track any type of stat from Tours to domain traffic. Even if you don't have something I'm interested in talking. email me at shap at twistys dot com.

[Trixxxia]

Originally Posted by shap

See I don't think it is 1-2%. I think with really detailed information the increase would be 7 to 12%. That is huge. You show your affiliates that type of increase and they'll increase traffic. And that 7 to 12% becomes a 15 to 20% in the long run. That's alot of $$$$$

Just a point I want to make is that yes, you can use the data to tweak what you already have out there, but it doesn't mean that you can use it on future projects first, wait for the results to show, then go back and cleanup or knitpick the first sites/galleries/tools to optimize them.

I will also agree that sometimes, too much info just puts a smokescreen on things and the focus is lost. Numbers are infinite, adding and extracting data from it is an ever growing and then eventually, you again lose the essence of the question/answer you were looking for in the first place. At that point, I call it a night and rest the brain to tackle it again in the morning

[milambur]

all I need is that the sponsor give me my 8 char tracking code, timestamp, amount, site/tour, referrer for each transaction that change my balance with them. When I feed that into my system magic happens. Just wish more sponsors would support it.

[Trixxxia]

Originally Posted by shap

By the way, if there are any stats freaks or programmers out there with existing products or products in production drop me a line. I'm not interested in an affiliate program. But plug ins or things to track any type of stat from Tours to domain traffic. Even if you don't have something I'm interested in talking. email me at shap at twistys dot com.

The one product I was impressed with was Rick Latona's Dollars.com - that thing had info out the wazoo It didn't catch on too much but I'm sure, had that been launched on the mainstream side and without any bugs (not sure if there were any) a few huge marketing companies would have been in a bidding war to get it for themselves.

[milambur]

I did not like dollars.com, think their focus was off.

[Shap]

Originally Posted by TopBucksTrixxxia

The one product I was impressed with was Rick Latona's Dollars.com - that thing had info out the wazoo It didn't catch on too much but I'm sure, had that been launched on the mainstream side and without any bugs (not sure if there were any) a few huge marketing companies would have been in a bidding war to get it for themselves.

Wasn't one of the problems with Dollars.com that Rick controlled all the data?

[ElvisManson]

Originally Posted by ElvisManson

if you are looking for more stats check out these guys.

They have some decent products.

http://www.omniture.com/products/web_analytics

[GigoloMason]

Keep in mind that there are probally a lot of people that do track stats religiously who aren't likely to advertise the fact. If you've got something that sells why go giving your research away to everyone out there to copy.

Shit I know if I found the magic formula I'd keep my mouth shut about it. ;)

[WiredGuy]

Originally Posted by ElvisManson

if you are looking for more stats check out these guys.

They have some decent products.

http://www.omniture.com/products/web_analytics

I wonder where you heard about this from
WG

[Pornopat]

Originally Posted by corvette

i agree

i.e. form hits/submissions/approvals are stats that (in my experience) do not get looked at enough

many clients just look at the # of sales

Corvette how about allowing affiliates to see declined sales and accepted sales by ccbill?
I would love to see those stats....

[SomeCreep]

Stats are important, but I'm no stats freak. As was touched upon by DrinkingHard, I'd rather build something new and see how it does, than spend time tweaking all my old stuff for a small % increase in traffic/sales.

Oh, and one thing about tweaking old stuff, is sometimes you tweak stuff the wrong way and actually break something that wasnt already broken.

[jayeff]

I studied statistics (at the London School of Economics) and started working life as a chartered accountant: thus I have a lot of faith in statistics. But I am also aware of their limitations.

Statistics which only assign a low probability to an event, or which have to include very broad parameters to be more definite, are almost useless for everyday business. The value of statistics is especially limited in this industry because we fail to sell so much more often that we are successful, therefore most of us don't have big enough numbers to do very much with at all.

Bigger sponsors are an exception, but even they can only dig a little way before their numbers will also become suspect. And the problem which even they face, is how to accurately interpret the figures they can collate.

A simple example, you have three tours and you want to know which one works best. It seems obvious: which generates the most sales per visitor? Not so.

Let's say that initially you had one tour, so for however long, all your affiliates pointed their surfers at that tour. Thousands of neglected pages are scattered around the 'Net, still sending visitors to it. Your more recent tours are only being used by newer affiliates and those who keep their sites updated, all assuming new is better. Mixed in will be a small number of affiliates who have actually tested which tours work best for them. Etc. The point is that the three tours do not get the same quality of traffic and therefore a simple calculation is not good enough. You are not really asking which converts surfers best, but which one adds most to the intrinsic conversion potential of the surfers it receives.

Framing questions correctly is where the problems start. Next comes recognizing which figures need to be modified before they will be useful. Then finally, as in the "simple" example above, there is the often insurmountable problem of how to generate those modifiers.

I suspect that between the nature of this industry's statistics and the likelihood that few people understand statistical theory, we probably end up with more wrong decisions than right ones being made as a result of people looking too closely at their numbers.

[Shap]

Originally Posted by Pornopat

Corvette how about allowing affiliates to see declined sales and accepted sales by ccbill?
I would love to see those stats....

Those stats are alot more stable than people think.

[kristin]

Shap, meet Steve Lightspeed. Possibly the biggest stats junkie I know.

[Pornopat]

Originally Posted by shap

Those stats are alot more stable than people think.

I would love to be able to have a look at my own stats as an affiliate.
Its weird that paysite owners are allowed to check them at ccbill where affiliates are not.
If they are stable or not....both ways they will give me valuable information.

[Trixxxia]

Originally Posted by shap

Wasn't one of the problems with Dollars.com that Rick controlled all the data?

Yup - one of the 'flaws' *if we can call it that* with it but you can see where a Marketing company in mainstream can use that for their clients

In adult, if it was sold as a module and priced to sell - it would have worked (meaning Dollars would have been welcomed better & sold) if everyone could have access to & control their own data (locally) without anyone else having access to it since it would serve each company & affiliates more to know exactly what link/pic/color/text was being seen/clicked/how many times etc...in order to better their product rather than rate how they stood compared to their peers & perhaps affect the amount of visibility they got in the end.

[Shap]

Originally Posted by TopBucksTrixxxia

Yup - one of the 'flaws' *if we can call it that* with it but you can see where a Marketing company in mainstream can use that for their clients

In adult, if it was sold as a module and priced to sell - it would have worked (meaning Dollars would have been welcomed better & sold) if everyone could have access to & control their own data (locally) without anyone else having access to it since it would serve each company & affiliates more to know exactly what link/pic/color/text was being seen/clicked/how many times etc...in order to better their product rather than rate how they stood compared to their peers & perhaps affect the amount of visibility they got in the end.

Bingo. I'm surprised he shut it down and didn't try to recoup by packaging it to sell.

Another short coming may have been him stepping on Evan's toes patent wise.

[seeric]

i've developed my own ways to decipher all the info at the processor and affil stats levels. site logs too. i think everyone has pretty much learned how to analyze their networks at the diff levels. its something you just have to sit and let your brain melt into for hours and hours. sometimes on saturday or soundays when i am jsut sitting around, i will intend to just look at one thing and then end up on some crazy 4 hour stats session.