Big Data, Psychological Profiling and the Future of Digital Marketing | Sandra Matz

Sandra Matz; Digital Psychometrics and its Future Effects on Technology; 34 slides.


  • Sandra Matz; Digital Psychometrics and its Future Effects on Technology; Keynote at ApacheCon; 2017-05-16; video: 23:08.
  • Sandra Matz; Big Data, Psychological Profiling and the Future of Digital Marketing; President’s Lecture, at The Berlin School; On YouTube; 2017-02-20; video: 1:10:52.


  • Cambridge Analytica
  • Apply Magic Sauce, Prediction API
  • myPersonality Project
    • myPersonality Database


  • Personality (Big Five, OCEAN)
  • Values
  • Life Satisfaction
  • Impulsivity
  • Openness to experience
  • Conscientiousness
  • Extraversion
  • Agreeableness
  • Neuroticism





Is Facebook Targeting Ads at Sad Teens?


Michael Reilly

      ; In

MIT Technology Review

      ; 2017-05-01.

The social network appears to leverage sensitive user data to aim ads at teenagers who say they feel “anxious” and “worthless.”

Concurrence on the Difficulty of Marketing the Chevrolet Volt

<quote ref=”here“>

What about TV ads? I don’t see Volt TV ads these days.
Norwicki: Generally speaking, the category isn’t advertised on TV. You go where the target customer for your vehicle is. And oftentimes people that are drawn to specific categories of cars, alternative-fuel vehicles in particular — those people do not view TV. They are online. They’re in social media. But they are not typical TV watchers. So just because you don’t see us on TV doesn’t mean we’re not advertising online and in social media.
If you advertise on TV, you’ll increase awareness, but that doesn’t necessarily mean you’ll increase consideration. So, by targeting we can more efficiently use our marketing funds.


GM’s statements totally align with my buying experience; I indicated such on the post-buy survey forms, and I’m sure many like me did as well.  TV didn’t help, wouldn’t have helped, couldn’t have helped.  The only appointment TV that I watch at this point is MLB and NFL, and typically that is 20 min or more out of phase to skip the commercials.  They [GM] advertise trucks on NFL.  I get that.  I already own a truck.  I’m not in market for a truck [yet]. My kids watch OTT-delivered video; e.g. Netflix or trawled shows in syndication on the TiVo.

What did help in the buy cycle?  Q&A from friends & colleagues in my trade.  These are tech-types who had already gone down the path, walked the walk & gone face-to-face with the “New GM” dealer network. I asked them about their experience: driving across the S.F. Bay, did they have to charge at work, did they charge at home, was it just “trading gas for electrons” or was there something more, etc.?  How did it run when out of juice?  As well, I learned that they had mitigated their ownership risk with a lease. Yet I wanted to buy for various reasons. The factory web outreach info was vastly helpful; the forums less so because of the UX, the unstructured conversations, attitude & chaos of the venues. Of course, the vehicle configurations one could construct in Build-A-Volt were not available at any dealer, but that bait-and-switch is true of any vehicle.  Build-a-Volt was great though because it familiarized me with the vehicle. Seeing the vehicle is important.

My real  persistent Single Nagging Question was: will the power plant “work” across time & distance at the same level of reliability that the Avalanche’s does: for 98% driving around town and freeway <= 40 miles to & from work. But-And-Also, I didn’t want to have to rent a “real” car to go to LA or on vacation, or to take my son to football league games in the north bay [Vallejo, Sacramento]; or worse have to rent a “real” car to go out to dinner with customers after work because my e-car didn’t have enough range.  Reliability over time is, of course, unknowable, but reputation precedes. Yet, the Volt chief engineer Andrew Farah’s testimonial that he drove Pike’s Peak in the car went a very long way towards my imagining that the car wasn’t another toy.  Oddly, so was Bob Lutz’s one-liner, something about Chicago and picking up a family member at O’Hare [cite].  But can I drive it to SF-LA — how will it handle Grapevine uphill after a 6 hr straight shot from Silicon Valley is my version of that.  I haven’t attempted that yet, but it does seem within the realm of convenient feasibility. Marketing is difficult, especially of “new” or “experimental” products; even with headwinds that drive prospective customers away

One way to address the marketing issue is to approach it on a cost basis, with incentives. The state & federal subsidy money was fun, but not a top-tier motivator. Deliberate minds know that recovering abstract incentives like those post-transaction are 1+ year out with substantial execution risk to capture it at tax time or filling out after market subsidy application forms, which might or might not fail on arcane bureaucratic grounds. I bought an L2, and had it installed by a contractor.  You need an L2 at home that you own and you control.  It’s a cost. Also, I’ve bought enough cars to know that what what you pay to drive off the lot is 150% the sticker price on the lot.  That’s how the system works.  As a consumer, you modulate this dealer markup by purchasing fewer vehicles and keeping them for longer.

I have to say that the surrounding culture of electric cars is not actually an attraction or strength in selling the concept.  Not for me.  I wouldn’t buy a car to get into hissy fits at my work with other employees about who is parked where or whether my car needs to be moved because there are too few electrical outlets.  All that does is broadcast to me that the owner has poor planning & buying skills to allow themselves to become dependent upon the kindness of strangers like that.  I bought a Volt so I could have freedom; the same freedom I have had with every other car & truck I’ve owned.  I can come and go when I please.  I get to park in the back of the lot and nobody tells me to move my car.  I would not buy a car to broadcast sanctimony or to whine at others about their lifestyles or choices.  So the “ICED OUT” entitlement, on-high national policy commentariat or intellectual pseudoviolence at abusers of parking norms & signage by electric car owners speaks to me as juvenile & extreme, continuing to define the electric car genre as fully-fanatical and still pre-early adopter; filled with wild-eyed crunchy types. The red-state/blue-state color of the discourse as well.  Why would I want to be a part of that?  Bizzarrely, from a marketing perspective, there is a still a continued self-loathing hypothetical line of grousing about the componentry on the Volt itself (fascias, mirrors, dashboards, etc.), which is often framed “in contrast with other $50K cars.”  Fair, but not fun.  It’s still an experimental vehicle choice; definitely not yet mainstream.  Marketing it sees it as on the cusp of cool in some areas, but definitely not yet “Crossing the Chasm” except here in the Valley of Heart’s Delight, even if Volts, Teslas and Leafs are de rigeur here.

I am mid-funnel in market in late-2014 through 2016 for:

  • a Suburban or Avalanche scale vehicle with a Voltec-type power train.
  • a Cadillac ELR, or ELR-V (whatever that is … I’m imagining it’s the “midlife crisis” variant for folks just like me).

Original Sources

in archaeological order, cut & paste derivatives on top, original works lower down.

Selecting Eff ective Means to Any End: Futures and Ethics of Persuasion Pro filing | Kaptein, Eckles

Maurits Kaptein, Dean Eckles; Selecting E ective Means to Any End: Futures and Ethics of Persuasion Pro filing; In Persuasive Technology; Lecture Notes in Computer Science; Volume 6137, 2010; pages 82-93 (12 pages); landing.


Interactive persuasive technologies can and do adapt to individuals. Existing systems identify and adapt to user preferences within a specifi c domain: e.g., a music recommender system adapts its recommended songs to user preferences. This paper is concerned with adaptive persuasive systems that adapt to individual di fferences in the effectiveness of particular means, rather than selecting di fferent ends. We give special attention to systems that implement persuasion profiling, adapting to individual differences, in the e ffects of influence strategies. We argue that these systems are worth separate consideration and raise unique ethical issues for two reasons:

  1. their end-independence implies that systems trained in one context can be used in other, unexpected contexts and
  2. they do not rely on, and are generally disadvantaged by, disclosing that they are adapting to individual di fferences.

We use examples of these systems to illustrate some ethically and practically challenging futures that these characteristics make possible.


  1. Introduction
  2. Adaptive persuasive technologies
    • Ends and means adaptation
  3. Influence Strategies and implementations
  4. Persuasion profiles
  5. Consequences of means-adaptation
    1. End-independence
      • Implications of end-independence
    2. Non-disclosure
      • Implications of non-disclosure
  6. Ethical Considerations
  7. Limitations
  8. Conclusions