Affiliate Survey – Player Value

[ThePOGG]

We all work in an industry where we have to blindly trust the statistics that are given to us by the programs we choose to partner with and I don’t know an affiliates that hasn’t at one point or another had concerns about the ROI they are receiving from some of their partners. The problem is that each of us individually – even for the biggest affiliates – has too small a sample size to draw valid conclusions.

ThePOGG Network is looking to counterbalance this by compiling a large scale survey of the value affiliates are receiving from the programs they work with.

Historically the biggest barriers to this happening have been that affiliates are reluctant to share performance information with their competitors. That is understandable, we wouldn’t want to do that either. As such we’ve put together a system that will anonymise the information submitted so that even we won’t be able to tell anything specific about your figures.

To this end we are looking for affiliates to take part in a survey.




How it works?

Selection Criteria

Firstly, we don’t just want any old affiliate contributing to this project. Allowing that would significantly devalue the data collected. There are two main groups we want to restrict from the data gathering process:

1) Operator representatives – for obvious reasons any person actively in the employ of an affiliate program has an active incentive to try to make player value appear higher than it actually is.

2) Unreasonably small affiliates – any site with low traffic or that only works with a small number of brands will produce data that is unreliable.

For the above reasons, any affiliate that wishes to take part will have to be vetted before they will be included in the survey. The factors that will come into consideration are as follows:

- Amount of unique content on the site
- Range of programs the affiliate works with
- Alexa traffic rating
Only sites which we believe to be genuine well established affiliate sites with a reasonable array of promoted programs will qualify for consideration. There will be other criteria but we don’t want to publicly discuss this as we don’t want to signpost how undesirable affiliates can get round our procedures.



The Survey

There are two aspects to the survey itself:

1) An approximation of the total traffic your site receives:

0-10000 unique visits/month
10001-25000 unique visits/month
25001-50000 unique visits/month
50001-100000 unique visits/month
100001+ unique visits/month

2) A rating for each program you work with.

The rating you award a program will be decided based on the average player value you have received from that specific program compared with the average player value you receive across your site/network as a whole.

Personally for this purpose I’d suggest StatsRemote, where I can look at the stats for the entire year and at the bottom I can see the average player value for our entire network over all programs and beside each program I have the average player value for that program. The rating should be established as follows:

5 - Average program player value is 50% or more higher than average network player value
4 - Average program player value is over 10% higher than average and less than 50% higher than the average network player value
3 – Within 10% of Average
2 - Average program player value is more than 10% less than average and less than 50% less than the average network player value
1 - Average program player value is 50% or more less than average network player value

Participating affiliates will be provided with a spreadsheet of affiliate programs and should fill in a rating for each program that they have had more than 10NDPs at this year.





How the results will be used?

Once the survey has been returned each result will then have a weighting applied to them based on the approximate traffic level listed in the survey (obviously a site seeing 100k unique clicks per month is more likely to produce more reliable information than a site with 5k unique clicks per month).

At this point all surveys will be collated and the scores for each program will be averaged. This will allow us to produce a single figure between 1 and 5 for each program to indicate the value that affiliates industry-wide receive from that program that will be far more comprehensive than any individual affiliate could make on their own.





How will the data be shared?

After the data has all been process we will compile a report that will be published on ThePOGG.com. We will also email a copy of the report out to all participating affiliates.

We will not publish results for any program where we receive less than 10 responses. For each program that we do publish a figure for, we will also publish the total number of respondents that were used to derive the figure.

We will not EVER publish or share the individual responses of ANY affiliate.




How do I take part?

Applying to participate is as simple as emailing me at webmaster@thepogg.com with details of the sites you represent for verification purposes. Please use a domain specific email address when contacting me. I’ll then review your sites and confirm whether or not you are eligible to take part.

Please note that the survey will be issued at the end of December and we will NOT go ahead with this unless we have at least 100 acceptable applicants.



Comments and feedback welcomed

 

[Guard Dog]

I would be interested, also, in having that published here at AGD if possible. Maybe a specific forum or in the general as a sticky thread. Sounds like a fantastic idea, though.

 

[ThePOGG]

Hey GD,

I'm more than happy to share this information - as a premise I think it's too valuable to affiliates to deal with in a selfish manner and if successful would be happy to see as many relevant sites publish the information as possible.

On that note - if this is successful, and depending on the workload, I would be inclined to repeat the exercise every 6-12 months.

TP

 

[ThePOGG]

Beyond the core group of affiliates that I spoke to before announcing this survey (CM, AGD and ourselves) we've had 1 affiliate sign-up. Unless there's an 180 degree about face in this trend we're going to get nowhere near the 100 affiliates required to go ahead with this.

Assuming that we're not going to hit that target, I will move forward if the signed up parties are still interested but will not be publishing the resulting information and will only provide it to the participants.

TP

 

[Guard Dog]

That is a bummer If I had the traffic required to give you any information at all I certainly would sign up.

 

[baldidiot]

For us part of the problem is I wouldn't even know how to answer the survey. We don't have one site, we have multiple ones. And there's massive variation between them - Brand A may be worth 50% more than average from Site A but 50% less than average on Site B. It often depends what the person is looking for and how well that brand fits the users needs.

Plus there's differences in vertical. Bingo players at Brand A may be worth more than average due to decent cross promotion, but sports players may be worth less due to crappy promotions or odds. But both accounts could have some from the same site.

I also disagree with the idea that a site with 100k visitors is more accurate than one with 5k. If that 100k is a tips site and the 5k is a review site, it could well be that it's the latter that sends more accounts. Surely accuracy is based more on RMP's rather than eyes on the page?

We could give a rough eye guide to how well we think the brand fares, but it'd be near impossible to quantify it in terms of x%. We'd probably only be able to do 'below average', 'average' and 'above average'. Unless you want to start breaking it down into verticals, countries etc...

 

[baldidiot]

On a side note I hope you don't think I'm bashing the idea in any way, would be cracking data to have. I just don't know how feasible it would be to get the data for (unless you're only looking at affiliates with one site) or if it would be that reliable.

 

[dfiocch]

OUCH! Just saw this...

I will compile the survey tomorrow (time for dinner here ).

Sorry!

Edit: mmmm... "old affiliate", you mean what?

 

[RyanWeb]

For us part of the problem is I wouldn't even know how to answer the survey. We don't have one site, we have multiple ones. And there's massive variation between them - Brand A may be worth 50% more than average from Site A but 50% less than average on Site B. It often depends what the person is looking for and how well that brand fits the users needs.

I would have the same issue.. I have a network of several websites (more then a handful), and I don't collect data on each of them individually. I usually judge the performance of a program and site by signups vs deposits, and not really website hits..

 

[ThePOGG]

Hey BI and RW,


No offence taken - feedback is appreciated.


To answer your questions:


The variation between the different values from different sites is actually part of what we're looking to see here. If you have site A, B and C and the same program performs above average at A, average at B and below average at C, that's the same as 3 different affiliates giving an above average, average and below average rating. You're just compiling part of the average yourself rather than submitting 3 separate surveys.

The above said, if you have separate affiliate accounts for sites and run Stats Remote independently for each site I'd be happy to have a chat with you and consider looking at a survey for each site independently if you'd prefer.

Regarding the differences between verticals, yes programs will deliver differing player values naturally for different verticals, but again we're comparing your average value with your average program value. The only point where this would become problematic would be if you're promoting 2 verticals with radically different average player values, where the one with the naturally higher value would be slightly inflated and the one with the naturally lower value would be slightly disadvantaged. However, with enough participants, the larger sample size would fade this sort of anomaly.

With respect to the traffic size - you're absolutely right that size isn't necessarily everything. I see gambling ads on eBay every day. They will have huge traffic but as they've also got traffic that is largely non-gambling, it's likely that they'd see lower player value than a site like our own and see far lower conversion. However, generally the affiliates that are likely to participate are likely to be more gambling targeted. While there will be higher traffic sites with weak conversion, again the law of large numbers would be expected to fade this issue with a larger sample size.

You’re also right that RMPs would be a more effective data point to use than traffic volume, but the intention is to ask affiliates for the least threatening data possible to encourage larger level participation. My guess here is that affiliates would rather give out a rough estimate of their traffic size than a rough estimate of the number of players they’ve sent to a program.

I do want to stress that the idea behind this is NOT to give a concrete, unchallengeable data point, it's intended to give a rough guide. We couldn't reach that strong a conclusion without out far more accurate data sharing and part of the point of this exercise would be to allow affiliates to share a level of data that they're comfortable with. If the exercise were to run successfully and were to be repeated and trends emerged over several surveys that would be stronger, but for the time being the purpose is simply to give affiliates a rough idea of whether other affiliates are seeing the same results with a program that they are.

The key to the value of this is that the more affiliates that participate the more. Even if there’s only 5 participants, each will end up with better information than they had with only their own figures, but as the numbers climb the reliability of the data increases significantly.

TP


P.S. dfiocch – the survey’s not been issued yet. That’ll happen around the end of the year. If you’re happy to take part drop me an email with your details so I can add you to the list!!