What is Duplicate Supression?

[webanalysissolutions]

As you can see from this part of the forum there have been multiple issues with Income Access based advertisers for years now, which has personally lead us to drop or minmise our exposure to anyone based on this affiliate software. I keep seeing people complaining about "duplicate supression", and although I get the general idea about what it is, and generally it seems a bad thing for affiliates, I wanted to start a thread to clarify the issue and get all the facts out in the open, either from affiliate experiences and/or from Income Access representatives.

From what I understand, the advertiser sends a file of stats to Income Access to update each affiliate's stats. However, prior to the file being sent, each day, or each month, referrals and data are stripped out of the file before it is sent to Income Access, meaning that affiliates lose valid referals, revenue, and tracking is therefore uncertain as the logic being applied to remove data is not transparent and understood by the affiliate.

So:

• What is duplicate supression?
• Is it only an Income Access issue?
• When does duplicate suppression occur?
• What logic determines a 'duplicate'?
• Which advertisers apply duplicate supression?
• What experiences have you encountered due to duplicate supression - have you lost referrals, not had referrals tracked?
• How widespread is duplicate supression, and how many players do you estimate are not counted/removed from the stats?

And anything else you might want to add.

If we, as affiliates can understand the full impact of duplicate supression then perhaps we can do something about it, by asking advertisers to modofy their logic to make it fairer and not discount valid referrals, or if it's something we can effect on our side we can stop sending traffic when we know it will be counted under the duplicate supresion logic e.g. if it's something to do with multiple registrations or clicks from the same IP then we can affect this.

A quick search of the GPWA forum will pull up a lot of posts mentioning duplicate supression, such as this one.
https://www.gpwa.org/forum/stats-iss...ate+supression

[Sausages]

• What experiences have you encountered due to duplicate supression - have you lost referrals, not had referrals tracked?

I don't think anyone will know for certain, most people put all of their trust in the program to be honest. If it wasn't for people like Shay, a lot would be ignored.

[PROFRBcom]

^^ Yea, I think that is the big problem. How can we know what is being stripped as affiliates? We need AMs (or former AMs) to step up and spill the beans on what is going on in these programs. Without a whistle blower, the task is nigh unsolvable... which is probably why it persists. I know that from experience we lose anywhere from 10-15% of the people we PERSONALLY walk through signups at various programs via one "glitch" or another. I can only imagine that the percentage is higher for those signups who come through our website and never actually interact with anyone on staff to walk them through account creation step by step. That adds up to a lot of lost revenue for us, and a lot of free revenue for the programs....

[webanalysissolutions]

^^ Yea, I think that is the big problem. How can we know what is being stripped as affiliates? We need AMs (or former AMs) to step up and spill the beans on what is going on in these programs. Without a whistle blower, the task is nigh unsolvable... which is probably why it persists. I know that from experience we lose anywhere from 10-15% of the people we PERSONALLY walk through signups at various programs via one "glitch" or another. I can only imagine that the percentage is higher for those signups who come through our website and never actually interact with anyone on staff to walk them through account creation step by step. That adds up to a lot of lost revenue for us, and a lot of free revenue for the programs....

So effectively "duplicate supression" is a semi-legitmate excuse for skimming in other words!?

I sort of knew that already, but it would be interesting to get down to, or just throw out some ideas (if IA or the avdertisers won't tell us their logic) as to when a player falls into this trap and gets wiped from your stats. If we knew this, or had a good idea about when it happened, we could change the way we refer clicks and limit our exposure to this.

My basic theory about referring players from the same IP for example - many people visit through proxy servers, hence may register under the same IP address. If duplicate supression kicks in here and doesn't track any new player that comes from an IP address that has previously had a player registered from, then I would definitely avoid sending proxy traffic to Income Access advertisers.

I have no idea what the logic is, but I do know that enough people have not had registrations tracked, so it would be interesting to learn the logic in order to work more efficiently with the IA advertisers we do work with on a limited basis.

Why can't these guys be just like the good advertisers like bet365, where you can trust in the stats and know that nothing is filtered, skimmed, altered, modified, or massaged?

[webanalysissolutions]

I know that from experience we lose anywhere from 10-15% of the people we PERSONALLY walk through signups at various programs via one "glitch" or another. I can only imagine that the percentage is higher for those signups who come through our website and never actually interact with anyone on staff to walk them through account creation step by step. That adds up to a lot of lost revenue for us, and a lot of free revenue for the programs....

Why do you still advertise companies on Income Access applying duplicate supression if you're being skimmed so much? Isn't it better to send the traffic elsewhere?

[TheGooner]

AS I understand it the 3rd party "duplicate suppression" software is run over the raw stats BEFORE the file is sent to Income Access for processing - so this activity is very difficult to detect or quantify from the affiliate end.

We have never been given a full technical description of exactly what duplicates are being detected and suppressed.
We have never been told how this file editing is audited and monitored. If at all.
- Is it actual players with more than one account? Via email address or Ip address? That would seem reasonable.
- Or is it something else that just allows a player to be untagged at will without documented evidence?

The only "evidence" I have on this is statistical - comparing the performance of programs with Income Access software and without Income Access software - and ranking programs accordingly.

We have 20 programs that we partner with - 6 of whom have IA affiliate software. They are ranked 9th, 11th, 16th, 18th, 19th and 20th.
These are household names, BetFred, BetVictor, Gala/Coral, Centrebet, SKyBet etc - their performance is awful.

By comparison programs like Bovada, Bet365, Ladbrokes, Rewards Affiliates and even newly acquired (for us) Mainstreet return 5-10x, and up to 20x more than the large brands running on IA software. There is not ONE brand running on IA software that I consider is performing to expectations.

This is just one affiliates opinion - but the empirical evidence is damning. All programs are advertised, promoted and linked in a similar manner on our single site - and exposed to 350K sessions (up to 1.5m pages) every month. The lagging performance of programs on IA software is a real issue.

Whether it is "duplicate suppression", poor cookie management, ineffective tagging or mobile customers, or another issue I cannot tell. What I CAN see is that I have yet to see a program running IA software that performs to the levels of programs that do not use the software.

Good luck in collecting and collating opinions - but getting hard data will be difficult as we do not see the before and after versions of data. All we can really measure is the absence of performing accounts.

[mojo]

All we can really measure is the absence of performing accounts.

There is one thing you can do for hard proof. Testers. Have other affiliates or players you know sign up and follow them through. I think APCW had started to do that but it should also be done regularly by affiliates over time.

From experience, there was a MGS program (not on IA) that we tested. We found that my player, who I am testing with, was showing exactly half of what he deposited. He would dep $100 and only $50 would show in my stats. We tested over several deposits and got the same result every time. That MGS program is now out of business.

Testing is the one tool for affiliates that tells the truth and I still test from time to time.

[-Shay-]

There is one thing you can do for hard proof. Testers. Have other affiliates or players you know sign up and follow them through. I think APCW had started to do that but it should also be done regularly by affiliates over time.

From experience, there was a MGS program (not on IA) that we tested. We found that my player, who I am testing with, was showing exactly half of what he deposited. He would dep $100 and only $50 would show in my stats. We tested over several deposits and got the same result every time. That MGS program is now out of business.

Testing is the one tool for affiliates that tells the truth and I still test from time to time.

A member and an affiliate manager (of an Income Access affiliate using program) tested the effects of duplicate suppression. My understanding is that the AM could see both players in his stats but the affiliate saw nothing. Said another way, the duplicate players were filtered out. Then, my understanding is that communications broke down and the AM went silent on the matter.

[Louis - Income Access]

Hi Webanalysissolutions

Thanks for the post as I think it's important to clarify for anyone who skimmed a long thread and didn't get all the detail.

Income Access has nothing to do with "duplicate suppression" and we don't use that term, or anything related, anywhere in our software.

The term was brought up by one of our unmanaged clients in a discussion with an affiliate who ran a registration test and the test account was not tracked. After the aff manager investigated with his data warehouse team, it turns out the player account was blocked due to an anti-fraud measure they have in place. The affiliate asked the brand to please disable this software for affiliate traffic. As such, we're not really in a position to comment on "duplicate suppression" beyond this one case where it was brought up between a client and an affiliate.

To quickly summarize our position regarding affiliate tracking concerns:

1) We rigorously test our tracking before any client launches and before any changes to tracking or reporting are set live.
2) We have extensive documentation regarding best practices and dedicated account managers who work tirelessly to ensure that tracking, reporting, payments and any other functions of our system are used to their maximum effect for strong affiliate relationships.
3) We work harder than most networks and software providers to keep our clients' programs "affiliate friendly" (Andy from affiliate guard dog can attest to our efforts to bring brands around).
4) We provide many more fail-safes for cookie-based tracking than the majority of other solutions available.
5) In the rare event that a tracking issue is caused, even partially, by our software, this is escalated to an emergency priority 1 item in our internal ticketing system until a resolution is found.

Gooner, when you first brought these concerns up, I suggested some steps for us to scientifically test the tracking in an unbiased manner, as this anecdotal data warrants real testing. I still feel that this would be a worthwhile effort, as you've been a valuable affiliate for a number of our brands and the relationship is worth saving. You're a smart guy, and I know you know Nicky, Sarafina and others at Income Access are not the type of people to risk skimming affiliate traffic. We've been with this industry since its infancy and seen how brands that cheat get burned, and we're certainly not going to put the axe to the goose with the golden eggs at this point in the company's trajectory. I encourage you to consider this.

Have a great week everyone

BTW - excuse my absence in the forum I've been off enjoying the perks of Quebec paternity leave

[mojo]

A member and an affiliate manager (of an Income Access affiliate using program) tested the effects of duplicate suppression. My understanding is that the AM could see both players in his stats but the affiliate saw nothing. Said another way, the duplicate players were filtered out. Then, my understanding is that communications broke down and the AM went silent on the matter.

I must have missed that. Do you know what the response was from the program and did it go public?

Edit: I posted the same time as Louis.

[-Shay-]

I must have missed that. Do you know what the response was from the program and did it go public?

Edit: I posted the same time as Louis.

My understanding is that it was basically a "let me look into this" followed by radio silence and no solution...

[justbookies]

My understanding is that it was basically a "let me look into this" followed by radio silence and no solution...

Duplicate Suppression ONLY applies to Coral. It is their own software.
I was the one who did a couple of tests and the AM did a test and he could see they were all tracked by IA software. The next day none appeared in the stats.
He blamed their own "duplicate suppression software" which filters out duplicates.
However I still get emails encouraging me to deposit into the account that never appeared in my stats, so it was NOTHING AT ALL TO DO WITH FRAUD or DUPLICATES! The intention may have been to filter out duplicate accounts (people who already have an account). But the software was at that time (and nobody has bothered to say otherwise since) very very aggressive and filtering a lot more than duplicate accounts.
Let me also say, duplicate suppression software even if it worked (and Coral's almost certainly is more like skimming that any valid attempt to filter the stats) is not needed. Duplicate accounts are merged or closed by the accounts teams manually. So to use it solely on the affiliate stats is a total nonsense.
Also, Shay has hit the nail on the head time and time again, the Coral affiliate department do not communicate unless you ask a simple query. Their AM ignores anything that actually matters. Radio silence. Any other program would have addressed these issues head on and buried all this discussion a long time ago.
In my case it was exactly as Shay said - I'll get back to you and of course (despite lots of pestering) he never did and never does. There is a long history of this at Coral and with this particular AM before he joined Coral.
Now Coral also have month end "reconciliation software". Last month it shave £1370 off my earnings. I now have screenshots for this month (and if you promote Coral take screenshots of your data today). If Coral could just explain FULLY and precisely what is being shaved and in what proportion and why then all this supposition would possibly go away.

[PROFRBcom]

Wow justbookies, that seems super shady and worthy of exposure. Thanks for sharing.

[webanalysissolutions]

To quickly summarize our position regarding affiliate tracking concerns:

1) We rigorously test our tracking before any client launches and before any changes to tracking or reporting are set live.
2) We have extensive documentation regarding best practices and dedicated account managers who work tirelessly to ensure that tracking, reporting, payments and any other functions of our system are used to their maximum effect for strong affiliate relationships.
3) We work harder than most networks and software providers to keep our clients' programs "affiliate friendly" (Andy from affiliate guard dog can attest to our efforts to bring brands around).
4) We provide many more fail-safes for cookie-based tracking than the majority of other solutions available.
5) In the rare event that a tracking issue is caused, even partially, by our software, this is escalated to an emergency priority 1 item in our internal ticketing system until a resolution is found.

This is all well and good from the Income Access side, but if you allow advertisers using your system to fiddle around on their side and modify data before sending it to the Income Access system then no amount of rigour, testing, or no matter how brilliant you say your software is, affiliates will still lose out from data being stripped from the files uploaded to IA.

There's no point having great racking, fail safes etc, if the referral IS actually tracked, only to be stripped from the numbers when you get the revenue data sent back to you in the form of affiliate earnings.

[webanalysissolutions]

Duplicate Suppression ONLY applies to Coral

Interesting stuff, I thought it existed elsewhere too. Certainly a lot of people I talk to struggle to earn much at all with Coral when benchmarking stats against other competitive UK facing bookies. Perhaps this stripping of data is the reason for that.

We were hit for 6+ months by BetFred actually with the old "fraud" players excuse. We were hit with several thousand GBP of lost earnings.

Duplicate accounts are merged or closed by the accounts teams manually. So to use it solely on the affiliate stats is a total nonsense.

Totally agree with you there. Why is there a need to start stripping affiliate stats before they're sent to Income Access? Surely as a business you would strip, merge, disable accounts in your core system to keep things tidy, then the affiliate data export would just take valid accounts. In fact, try registering as a duplicate with people like Coral - a lot of the duplication logic is actually captured at registration, so it's surprising that any company would need to go stripping out duplicates.

Now Coral also have month end "reconciliation software". Last month it shave £1370 off my earnings.

Ouch. I have seen this far too often with other advertisers too. In April alone were were £1500 up with one Income Access based advertiser (not Coral) only for all the days' stats to be changed overnight later in the month. We finished the month only £170 up.

Thanks to everyone for their input. If anything it's given me more insight into this, even if we can't nail down the logic, or the size of the impact, it's definitely leading me further towards dropping Coral completely - something we've been considering as a business for a while now anyway due to poor conversion.

[webanalysissolutions]

The only "evidence" I have on this is statistical - comparing the performance of programs with Income Access software and without Income Access software - and ranking programs accordingly.

We have 20 programs that we partner with - 6 of whom have IA affiliate software. They are ranked 9th, 11th, 16th, 18th, 19th and 20th.
These are household names, BetFred, BetVictor, Gala/Coral, Centrebet, SKyBet etc - their performance is awful.

Can echo much of what you're experiencing too Gooner. Ultimately program performance dictates whether such advertisers get space on your site, if at all, but it would be great in an ideal world to work out the underlying reasons for why Income Access programs perform far worse than non-Income Access programs in both of our experiences. So many IA programs underperform relative to non-IA programs in our experience that it just doesn't seem coincidence.

[-Shay-]

I have been told that Income Access allows the "luxury" of providing their partners the opportunity to review their data before it is uploaded into the software. A review before submission certainly opens up to the possibility of manipulation, omission, and even addition of data.

The process should be automated and seamless - not subject to review and not subject to any type of filtering via duplicate suppression or whatever you want to call the filter. This is why I have more trust in programs who push real time/near real time data. In my mind, there's less time to catch the drop if there is no pale that accumulates throughout the day to sift through. I know that IA has an option for hourly updating...

[BonusBets82]

Great posts from everyone here and cheers to Gooner for the stats and to Justbookies and Shay for the additional data. Since I see some fellow affiliates providing some real good transparent data here, I'll chip in with my stats, as they are strikingly similar to Gooners. I'm talking 4-5 years of data from my sites - from 100,000 - 200,000 hits per month.

Performance by Program Software
1 - Not IA
2 - Not IA
3 - Not IA
4 - Not IA
5 - Not IA
6 - IA
7 - Not IA
8 - IA
9 - IA
10 - IA

4 out of 5 of my poorest performers all use Income Access I've taken into account conversions/100 here as well as overall return, as positions/promotions/visibility can vary obviously.

Also agree with Justbookies on the Coral aff manager. That "Hodgey" guy or whatever he's called seemed to vanish every time something other than a nice and easy 'positive' question was asked. Reminds me of that Luke guy from Paddy, don't think I've seen him since Paddy took a slagging for their awful program/session cookie however that's off topic as they use NetRefer. Of course, both guys seem to be well liked, I realise that but can't comment as I don't personally know them, but it doesn't make a bit of difference in my opinion, questions and huge concerns need answered asap.

Despite the long detailed posts from the IA guys and their visibility here on the forum, our worries and revenue losses from IA programs still exists.

[webanalysissolutions]

This is why I have more trust in programs who push real time/near real time data. In my mind, there's less time to catch the drop if there is no pale that accumulates throughout the day to sift through. I know that IA has an option for hourly updating...

Interesting thoughts. Affiliate programs have evolved over the years to learn that trust is one of the key selling points in attracting useful/big affiliates. Near real-time information would certainly be better than every 24 hours when it comes to transparency

[-Shay-]

Of our top ten, IA-using sites rank 5, 9, 10 with four more scattered in the top 20. We work with 25 brands in all. In like for like situations, IA brands underperform in earnings and conversions against brands using proprietary tracking and/or tracking programs such as NetRefer.

I might add that while I cannot physically see the leak, there does seem to be a puddle of water on the floor and that puddle seems smaller when I focus on the bucket of non-Income Access sites than when I focus on the IA bucket.

I may be blind and lack pure scientific data, but I know which bucket to focus on filling if I wish to collect and retain as much water as possible.