I’m specifically labeling this an Opinion as I’m going to come down rather hard on DoubleClick’s Performics division which, at the recent Affiliate Summit event, announced that it has proof that one’s affiliate marketing program (or “channel") is valuable (did anyone ask?).

Again, this study, like so many, asks retailers to set aside critical facts—i.e. the self-serving sample used.

I’ve said it before and I’ll say it again: Commissioned (under-written by parties with vested interests) research is bunk; always has been, always will be.  The latest offense to our intelligence comes in the form of a Performics-funded, comScore Networks executed study that looks across the surfing habits of around a million profiled (demographically) Web shoppers.  These shoppers allow comScore to track and log every single move they make across the Web in return for the rather dated, 1990s-esqe “faster Web connection speed” promise. 

Garbage In = Garbage Out
Without a good sample, research falls flat.  In the case of Performics’ study we know a good deal about the comScore sample (the consumers) but what about the other study participants… namely the affiliates or “publishers” as we now call them.  The study focused on 63 merchant sites with active affiliate marketing programs across all three major affiliate marketing networks and 19 “large affiliate publisher sites.” Nineteen?  Yup… nineteen and we have no idea who or what flavor they come in or why a more statistically representative number was not used.  This information is not provided although nothing could be more relevant!

(and able to impact/skew the data analysis)

Irrelevant Results
I suggest that, given the “secret sample,” we are left with one intelligent option: discard the findings as there is so much room for manipulation of the data.  Performics has given us a…

1) Limited (19 large affiliate publishers)
2) Partial (no information on who or what type)

... sample to accept as being worthwhile, honestly representative of one’s own affiliate program.  This simply cannot be the case as retailers are (or should be) much more discerning about their decision-making processes.  Who has 19 affiliates in their program and, more importantly, neglects to categorize them (indeed set customer acquisition, ROI expectations based on these categories) in this day and age?

UPDATE (see Comments): Performics reports, “The 19 publisher sites were an additional (emphisis added by editor) analysis of the study to provide Performics’ clients with demographic and performance data for those specific publisher sites.”

Propping Up Affiliate Marketing
Which brings me to my closing remark: Why is affiliate marketing in need of propping itself up?  What is wrong and why is this study needed… in particular in the retail segment?  I think the answer is as obvious as it is troubling and signals that we’re on the cusp of change.  Indeed, mergers and acquisitions seem in order.  Yet what about all the trouble points—from legal battles (Lands End... Dell) to behind-the-scenes technology battles (eBay) that drive marketers away from affiliate marketing networks?  It looks like the wagons are circling.  What’s next for affiliate marketing?

In announcing this study, Chris Henger, VP of affiliate marketing at Performics, said

“Our clients consistently meet their online marketing goals via effective affiliate marketing and our more than 300 advertisers recognize the high value customers this cost effective channel delivers. This research essentially provides the industry with a data-driven, third-party profile, of the affiliate consumer.”

I invite Mr. Henger to respond to my criticisms, consider providing reasoning behind the sample size (on the affiliate side) and release names of affiliate/publisher companies and/or categories so retailers can better understand this study.

Perhaps a Manifesto is needed.  I call on Sam Harrelson to put together something similar to SausageManifesto.com (a grass roots list of demands/considerations for, in this case, the search marketing industry’s leaders/power brokers to take the click fraud bull by the horns—stop the tapdancing).