- 6 octobre 2009

Comscore CEO Dr. Magid Abraham Accepts AMA's Coveted Parlin Award!

This morning, Comscore CEO and co-founder Dr. Magid Abraham accepted the AMA’s 64th annual Charles Coolidge Parlin Marketing Research Award. The Parlin Award is a preeminent national honor awarded to those who have demonstrated “outstanding leadership and sustained impact on advancing the evolving profession of marketing research over an extended period of time.” Past honorees have included such marketing industry luminaries as Peter Drucker, David Ogilvy and Philip Kotler.

Below is the full text of Dr. Abraham’s acceptance speech.

Good Morning!

I would like to thank the Parlin board of governors and the AMA for this great award. I am honored to be chosen and humbled to be counted in the same group as the illustrious past winners.

Let me say a few things about my background. I grew up in Lebanon on a fruit farm. The obvious question is what got me from loading apple crates on the backs of donkeys to a career in Marketing Research of all things. The answer about leaving the donkey line of work is simple: I had greater ambitions than looking at donkey’s behinds! Perhaps the more interesting question is: Why market research and not medicine, science, engineering, finance, law, or any of the myriad of professions more likely to be familiar to a farmer’s son?

Truth be told, I am surprised that this is what I ended up doing. I have always loved Math and Science and was particularly fond of physics. My stated ambition in my high school yearbook was to become a nuclear scientist. As I was about to graduate from college, there was a crisis in the nuclear industry in the aftermath of Three Mile Island. I was no longer sure that becoming a nuclear physicist would guarantee I would not go back to walking behind donkeys. I played it safe and decided to pursue graduate studies in Operations Research at MIT.

The business courses I took in my first year left me with the initial impression that Marketing was just an artsy folksy discipline that I should stay away from, and that the name “Marketing Science” was a stretch at best.

The turning point came when I invited Professor John D. C. Little, who is also a Parlin recipient, to give a seminar to the graduate Operations Research students. John presented his work on using multinomial Logit to predict consumer choice in the coffee category using scanner panel data which was a novelty at the time. I was fascinated that such data actually existed and was seduced by the mathematics involved. In fact, I found the modeling of choice probabilities a bit reminiscent of Quantum Physics. With my interest piqued, I took a summer job at Management Decision Systems, a marketing research company John co-founded. I fell in love and never looked back.

As a data geek, I quickly realized that good data is just as important as good models for answering research questions, if not more so. My first R&D assignment at MDS was to work with Professor Len Lodish from Wharton to develop solutions for evaluating trade promotions by CPG brands. The system we developed, Promoter, was created at a time when the norm for measuring retail sales was bimonthly audit data from AC Nielsen, which was totally inadequate for the task at hand. We ended up primarily using weekly shipment data disaggregated at the key account level.

But I soon realized that nirvana lurked in the shiny laser beams of supermarket scanners which collected sales transactions to consumers in real time. With weekly scanner data, the estimation of incremental sales due to promotions is far simpler, richer and more precise. It enabled us to measure the sales lift generated by different types of promotional tactics such as newspaper ads, end-aisle displays and temporary price reductions.

I quickly focused my energy to help in building a national scanner data source which became IRI’s Infoscan, and adapting the Promoter methodology to measure promotional elasticities on a massive scale, for every product, retail store and weekly period.

Then, in an industry first, we made the elasticity metrics routinely available with sales data. Clients did not need to run a special model to estimate those elasticities. They just read them in their market share reports. We also decomposed market share into two components: Base share, which is the market share a brand would get in the absence of promotions, and incremental share contributed by short term promotional activity. This new paradigm, a staple in the industry today, was very useful in separating the impact of short term tools that brand managers can use to increase sales, from the longer term drivers of brand sales including brand positioning, everyday pricing, advertising and new products.

Along the way, I came to appreciate the importance of research automation. To deliver base and incremental sales, we had to generate over 100 million baseline estimates per week. We simply could not have analysts review the estimation output. It was imperative to build a system that mimics what a smart analyst would do in handling special situations and data exceptions. To this day, people tease me about a 1985 Promoter promotional video where I proudly boasted, only as a 25 year old could, that “Promoter’s experienced eye cannot be fooled”. But that was no idle joke. I was very proud of “Promoter’s experienced eye”. It allowed us to have confidence in data estimated on billions of observations without human supervision.

Market Researchers are typically proud of their analytical acumen. My motto is: “Automate to Dominate”. Automate analytical intelligence to dominate mountains of data. Let silicon do everything that silicon can do, and save your brain cells for creative analysis and insightful communication.

It is almost always the case that end-users of market research information lack the interest, skills or time needed to fully leverage the information we provide them. We are more successful when we simplify, provide easy access, and deliver results in the most useable or actionable form.

I applied this principle in designing another expert system called Sales Partner. Sales organizations recognize that data and insights can be very powerful selling tools. At the same time, salespeople are not data geeks like some of us. They typically do not have the time or the inclination to analyze data and use it in a sales pitch. Sales Partner mimicked the thought process of a clever sales guy and automated the creation of cogent, data driven sales arguments to support a specific sales objective. The system became popular very quickly and was probably one of the most gratifying products I was involved in.

As CEO of Comscore I have many of the common CEO responsibilities. However, I am far from being a typical public company CEO. I am frequently happily engrossed in solving many of our core methodological and analytical problems, whether they relate to sampling and weighting methods, predictive models of consumer behavior, reach and frequency models, advertising response models, or, more recently, the development of hybrid measurement methodologies integrating audience projections from a sample of individual panelists with a census of usage measured by web servers.

This hybrid integration is akin to a measurement holy grail combining the benefits of rich granular information from a sample with the accuracy of a census. While conceptually simple, this turns out to be a challenging problem putting to the test the creativity and ingenuity of our best and brightest minds. Personally, I have enjoyed joining the fray and working through the conceptual and mathematical complexity of the solution.

This farmer’s son is fortunate to have had the opportunity to create scalable research data sources and models that are used daily by tens of thousands of users worldwide. But, at the core, he is happy to be, like you, a market researcher.

I think that people in this industry are a special breed. They believe in the power of data analysis which they use as their stock in trade. They frequently set out to exercise this power by starting new companies, and constantly searching for and solving new problems. They are obsessed with modeling consumer behavior, which can be more fickle and less predictable than subatomic particles. This is a profession that values creativity and entrepreneurship. A profession that spawned many pioneers who built the tools needed for marketing success in highly evolved and competitive markets. You certainly can be successful without good marketing research, but you are always more successful with it.

I would not be here without the help of many people who supported me throughout my career. People such as Gian Fulgoni, former CEO of IRI and chairman of Comscore, John Malec and Gerry Eskin, co-founders of IRI, Len Lodish my partner on many research projects, Glen Urban and John Little, my PhD dissertation advisor and mentor. I am also thankful for all the people at Comscore whose tireless efforts and ingenuity helped build the company to its current success. Finally, my wife, Linda Abraham, who is now Comscore’s CMO and was my co-founder at Paragren, who has always been a great sounding board and a reliable source of support and inspiration. I am proud to accept this great award on my and their behalf. Above all, I am especially proud to be part of this noble profession.

Thank you!

Plus