Oops, just saw your answer above. Please ignore this question.

“[3] I hasten to add that teaching is very different from public speaking and is emphatically not a performance.”

Can you elaborate on some of the differences a presentation that’s oriented towards teaching?

It seems to me that complex data mining simply isn’t needed for this purpose, and that to claim otherwise is an instance of “if you have a hammer”ism – if you’re a data miner, everything looks like it could be improved with data mining.

A concrete example: my local store, HEB, gives me a coupon for money off five Hot Poockets, when I buy four. When I buy eight, it suggests I buy ten. And so forth. It encourages me to scale up by about 20-25% on one or more products in my cart, each shopping trip. They don’t always offer coupons on hot pockets – it can be other items. they don’t always offer coupons on things I specifically bought, but rather on things that are often bought with them: if I buy hot dogs, they might offer me a coupon on hot dog buns even if I didn’t buy them. So while, this week, I might’ve bought my buns elsewhere, next week, I might buy them in HEB, too, making me more loyal.

They don’t need complex profiling: they just need to know that “you bought a few of X on this shopping trip, so are likely to buy it again, and with some money off, can probably encouraged to stretch your spending on that item a little further, and buy 25% more. Further, by giving you coupons you can actually use, for products you actually buy, we can make you more loyal.”

They don’t even need a store card system. A store card and rewards point system would get, I would argue, negligible benefit over and above this *delightfully anonymous*, point-of-sale coupon system.

Anonymity is nice, and I appreciate it as a consumer: tracking-by-store-card is a whole less nice, so please don’t encourage it, even if that is what your research pertains to.

Not so.

I am not at all surprised that this could be a viable technique. My point was that the evidence as presented on this page does not appear to strongly establish that viability.

That is what I wrote, that is what I meant.

he has an interesting chapter on “incentives” which I regard as key and which could easily merit/fill an entire book itself. he says we are possibly in the middle of a realignment, even paradigm shift, & agree with this somewhat. but its a long slow process. akin to turning a battleship around. one can maybe see an early example of some shift in “incentives” in the stackexchange “rep/point” model.

when I started my blog, thought that finding likeminded volunteers/collaborators/contributors might not be so hard. there are tens of thousands of open source projects eg on github or sourceforge. but how many open source/collaborative science projects are there? there seem to be very few. polymath is a great model, but it doesnt at all look at motivation and incentive, its sort of the elephant in the room, swept under the rug. and (gasp! full disclosure!) its not really as successful as you might think. it turns out to be very challenging even finding a dribble of people (or even a single other person!) to team up on semifamous problems widely regarded as significant, eg in my case the collatz conjecture, or work on famous open problems with large rewards, eg P vs NP.

I worked in a scientific lab 4 yrs as undergraduate (neurobiology) and saw the incentive system and decided against it in favor of “industry”. there seemed to be maybe only one person in the lab making good money– the head of the lab, who also seemed somewhat an anomaly. the very choice of science seems to necessarily/fundamentally involve a sacrifice of incentive. bringing this out in the open as you have done in your blog is part of the evolution toward something better.

there is a very interesting book on the subject of competition vs cooperation and evolutionary arms races, try “red queen race” by matt ridley. also the economic concepts related to “tragedy of the commons” seems to be relevant also.

worth a whole blog at least! intend to write further on this on my blog.

If only it were that simple. Doing science in the Wikipedia model where volunteers contribute in their spare time would take us back a couple hundred years to the amateur era of science. Modern science requires several years of training to be productive, and we want to do it on a large scale, so it needs to be done by professionals. Finding a way for people to collaborate openly while also getting paid (and hence, having a way to measure contributions fairly and accurately) is a phenomenally more difficult problem.

Alexandr

Yes, you definitely have found a way to buck the trend and seem to have found a really good balance! Don’t worry–the competition/pressure to publish isn’t the only thing making me reconsider a future in academia. I just don’t think it’s what I’m supposed to be doing in the future.

I hope you also take into account the positive message I expressed toward the end — while playing the game might be the safest way to succeed in academia, I don’t think it’s the only way. I’m quite happy with the balance I’ve found between publishing and doing research the way I want.

PS: Would really love to hear the reasons that changed your mind and made you decide to apply to (and then join) academia!

Krogers “gets” me: nearly every coupon they send is for a product I like and buy regularly. Krogers also knows how disorganized I am, and attaches coupons directly to my scan tag.

On the other hand, Krogers knows more about me than even Google or Facebook in a lot of ways … a little scary.

“I think you will find that at least a substantial minority of customers hate coupons”

I suppose we disagree on just how substantial that minority is.

“I would go so far as to say that price discrimination of any kind will always be perceived as unfair by all customers who are aware that it is, in fact, price discrimination.”

Quite possibly, but that doesn’t say much from my perspective. People hate anything with the dreaded d-word in it, justified or not, so of course if they see it as price discrimination they’ll hate it. As a corollary, take an example of price discrimination that’s seen as socially just — say need-based educational scholarships — and try convincing people that it is, in fact, price discrimination.

More to the point, I have a rather dim view of what it is that consumers do perceive. I went into this in some detail in my previous post — price fairness perceptions are highly susceptible to the way the transaction is framed.

Good point. I’d be curious to know if those are seen by retailers as a way to push new products or a way to price discriminate (or both).

I would go so far as to say that price discrimination of any kind is always perceived as unfair by customers who are aware that it is, in fact, price discrimination.

I would go so far as to say that price discrimination of any kind will always be perceived as unfair by all customers who are aware that it is, in fact, price discrimination.

Meh. I don’t need data, I can always generate my own data, the data isn’t the point of the paper…

No, but you can get the data by emailing us.

It is a long time since I sent my last comment/question :-)

Now that all major telco operators are starting to sell our profiles jointly with location data anonymized and aggregated I was wondering if you have research a bit on the issue

Any idea about anonymization combined with aggregation approaches they are using?

Dynamic or persistent data masking?

What can you tell us about data masking techniques are they really efficient?

Thank you so much for all your posts. Keep on