Chris Williams and I exchanged emails for this interview and it has been slightly condensed and edited
Marc Goldberg: Chris tell us a little about your journey
Chris Williams : My last role was at the Association of Canadian Advertisers ACA but I started working in digital media with a project for Apple in 1993. We used browser recognition to change the creative – if you were on a Mac the image was a little angel and for PC users we had a little devil – a cute one. The point was to welcome PC users that were shopping for Apple computers. We had no idea how to analyze the report, it was all page hits, site analytics was still in the future.
From there we did we a virtual trip down Route 66 for a beer company, wrote a banner that got an astounding 38% click rate (cta: “click here you little worm!”) and did a campaign for a Canadian competitor to nascent Amazon (“we bring books to Climax – Saskatchewan). Then 2001 hit and the wheels came off the bus.
MG: MMM Beer. When did you get interested in media technology?
CW: Havas developed this thing; Artemis, it took all the data out of the ad server so you could make custom reports. It was way ahead of its time and it forced you to think about what you can do with tags and pixels.
When I was at IAB, Julie Ford and I found the ad ops committee to be one of the most fun groups. We connected them with Steve Sullivan and Chris Mejia (IAB US) who did the deep dive on Bots and Invalid traffic. Chris recommended we join the Media Rating Council – which wasn’t in the budget, that was a missed opportunity that I regret.
ACA was committed to upholding strong standards and they joined MRC when I joined so then we were drinking from the fire hose of MRC audit reports.
MG: Is that when you started seeing The Dark Side of Digital Media?
CW: No, that started much earlier. A board member of IAB spent a day going through all the fraud and bot techniques, the cash out routes and who was making the money. Then there was ACA’s Bot Fraud report using a technology that was later picked up by Nielsen. I was writing the report and looking at this massive spreadsheet, some of the sources were 100% Bots, nobody was doing effective CPM calculations. It was simple just to go straight to LinkedIn and say this is what I am seeing and it kind of sucks. Then people started to send me things that they were seeing too. My favourite was “Supreme Traffic Bot” with all their advanced tutorials on Youtube, they’re still there too.
MG: It is not all the evil dark web and hidden underworlds, it is all in plain sight isn’t it?
CW: Exactly! I was watching this Cambridge Analytica develop and saw that the company they used; Aggregate IQ was based in Victoria BC so I gave them a call. I thought maybe they would do a webinar for ACA explaining the use of OCEAN psycho-profiling before they became famous. We had a long chat about “interesting projects” but no webinar. The other thing out of Victoria was the Youtube brand safety blow up on the creepy comments on kid’s videos. I phoned him too, poor guy was taking a lot of heat from other Youtubers who saw their revenue drop overnight.
It was about that time the report written by your CEO at MMI Shailin Dhar and other voices who were loud about ad fraud I leaned in. I think we connected on the issue of Inclusion lists and it was becoming clear that all of these issues of brand safety, attribution measurement and invalid traffic really needed the help of the Media Rating Council.
MG: Yes we met and I love that you really promoted a lot of this ugly side of the world to your members, not to scare, but to educate. I think it is the only way to discuss these problems…which are still….problems today
CW: Yeah for me 2019 was an epiphany. Chris Cable of Diageo presented at ICOM and in 8 minutes gave a succinct blow up of multi-touch attribution, privacy and brands while setting a foundation of the importance of geography in digital media. Same conference, Shailin did a deep dive on rendered impressions versus billed impressions and Tom Droulliard demonstrated digital publisher audits. Then Rick Bruner explained what Central Control is all about.
I came away from that conference thinking that Marketing Mix Modelling was the way but it had to pick up the speed and granularity of digital. Second, that MMM needs the basis of media planning which includes geo-separated control/exposed test aligned with the advertiser’s CRM outcomes, not conversion pixels. And finally if there was to be a positive advertiser benefit the impressions should be coming from quality media vendors displaying and billing off the most stringent MRC standard.
MG: So what are you working on now?
CW: As an independent contractor, one thing I am doing is helping out with the pre-seed startup. Winston, the founder has built out the evolution of MMM, plus strategic forecasting and media planning in one platform based on a synthetic population. It gets away from questions about identity and privacy because synthetic people equals no personal information. And it’s cross media and based on open source geo-cohorts.
MG: How does this address questions of trust and transparency?
CW: Google has proposed cohorts as part of FLoC and that is the right thing to do BUT… the question is who defines the cohorts, how are they defined and how interoperable are the cohorts? All media, all of it, resolves to geography in some way so geo-cohorts are the lingua franca bridging analogue and digital media. Geo-cohorts or postal codes are public, open source and don’t deprecate. No one wakes up one day to hear the Government has decided postal codes are no longer supported and everything is moving to IP addresses including your house. So that is fully transparent and postal codes have been around for 50 years, no data leaks yet, no one is selling lists of postal codes on the dark web.
Meanwhile conversion pixels are weak. They are a huge data leak from advertisers to the platforms – who knows what data is being used for, they are vulnerable to IVT, so it’s time to move on from them; media cohort / sales cohort – much more trustworthy and neutral to all media.
MG: But what about personalization?
CW: Personalization positions the brand subservient to the person, “this it all about you” and I think that’s incorrect. Brands need to be themselves and communicate very broadly which is the opposite of micro-targeting. Also – response or activation too is the opposite of branding. A strong brand can defend its margin without temporary price reductions.
The argument for personalization is usually made around media efficiency – cue the Wannamaker quote – but there is a difference between waste and spill. One day when I decide to get off the sofa and become a triathlete, my short list of athletic brands awaits my credit card. In the meantime, athletes benefit from my appreciation of the badge brands they wear. Brands are imprinted through moments of truth, when the promise is kept and there is a long look back window between sofa and track. I bought a motorcycle last fall – 27 year look back window to the day when their marketing sucked me in. 7 days? As if.
MG: A motorcycle? Join a biker gang too? Imagine a biker gang in Canada cruises around and performs good deeds rather than terrorizes folks.
CW: It’s just brands preying on middle age angst
MG: Well I am not your parent or wife but I will say I don’t approve on their behalf.