Is Amazon really that cool as we keep saying?

For all that buzz around Amazon’s sophisticated analytics and its targeted book recommendations, it is worth asking in Kevin Hillstrom’s priceless, heretic style: Is it just hype or does it really make the big difference for their business?

Do we really buy from Amazon because of the recommendations, personalized emails, the behavioral targeting through widgets?

Or do we buy from Amazon because they always have the cheapest price? (Not to the least because of the 3d party vendors and used books that are linked in.)

If they stopped being the least expensive would we still be buying from Amazon?

In other words, are they really competing on analytics? Or are they competing on price?

What is our willingness to pay extra for the kind of “marketing as a service” that Amazon has perfected?

Of course … this question isn’t really about Amazon, in the end. Much rather I am trying to double check what the true value of sophisticated analytics and targeted marketing are. All hype aside.

Separate things: What you will buy vs. where you will buy it

There is no doubt that Amazon is the go-to place for doing your research on books.

But being the greatest place for researching books doesn’t necessarily mean that people will buy the books there if they can get them cheaper elsewhere.

I imagine we all go to many web sites to research what car, electronics, gear, etc. we should buy. Where we will buy the item that we settle on tends to be a different question though. We might check on eBay or Craigslist, for example.

So are we giving too much credit to Amazon’s sophisticated analytics and marketing?

The answer …

As Anil Batra was joking yesterday when I saw him at the OMMA Metrics and Measurement in San Francisco, a typical consultant’s answer to such a question could be: “That depends on … what it depends on.”

It depends on …

Of course, I don’t have Amazon’s data. But I think that the answer will indeed differ by buyer segment:

  1. High value prospects that buy books frequently and in higher quantities will likely appreciate the convenience and time savings. But given that these people buy so many books they would also be the segment that could get the biggest total savings by being disloyal to Amazon.
  2. Infrequent buyers that are strapped for time will value the convenience over a few bucks of savings per book.
  3. Infrequent buyers that are strapped for money but not for time would be more likely to put in the extra 5 minutes for buying the book at the cheapest vendor. Most books aren’t big ticket items. So this effect will likely be much less pronounced than, say, with electronics. But it would be there to a certain degree.

In the end, maybe the answer has not all that much to do with frequency but is a function of

  • how much money vs. time buyers can save.
  • how much value the individual puts on money vs. time

Time is money and money can buy time

The more time Amazon can save its book researchers, the more of them would buy their books on Amazon (assuming price is fixed). Still, those buyers who value money a lot more than convenience and time would be the hardest, if not impossible, to keep.

Anything else they could do?

Barnes and Noble has (or had) essentially a frequent buyer card. You could pay an annual fee and would get x % discounts on anything you buy.

But that really is just a return to the strategy of competing on price.

Could Amazon withhold access to book research features unless a buyer … purchased something in the last 12 months or joined some kind of for-pay club? That would be a return to subscription based content. Seems like it would backfire badly.


It appears that the true value of all those analytics and targeted marketing for the retailer are in drawing the crowd into their (online) store. They get a shot at making sales (and cross-sales) that they wouldn’t otherwise have.

But converting researchers into buyers requires more than just targeted marketing. It also requires convenience, a “good enough” price, and of course customer satisfaction with previous transactions.

How would you go about measuring the value of targeted marketing effort XYZ?

If you have to measure ROI of targeted marketing effort XYZ you would probably do it through controlled testing.

That would be easy for Amazon’s targeted emails.

It would be harder for their book recommendations because you would wonder where they disappeared to if you fell into the control group.

Oh … but you could make them deliberately untargeted. Say you find that dinosaur book that the individual bought and shipped to somebody as a gift 3 years ago and recommend more dinosaur books. 😎

9 Comments on “Is Amazon really that cool as we keep saying?

  1. I can buy any item I want with about 3 clicks on my iphone in less than 30 seconds with zero hassle.

    If I try to buy a replacement power cord on ebay on my computer I’ve got to go thru 5 or 6 screens just to pay. There’s no way I’d even try that on my iphone.

    for me that’s important and cool 🙂

  2. A great post Akin. Even the best targeted marketing cannot create ‘needs’ in vacuum. The customers at least should have a minute seed of desire that then can be cultivated and exploited with a good targeted marketing program.

    As to Amazon, I have to admit I go to Amazon as a first leg of search/research on books, music etc. mainly because of two reasons: (a) the brand name (even though I know other booksellers have comparable sites, I still have Amazon stuck in my mind more deeply and so think of it first), and (b) convenience (being online it is much easier to search from anywhere/anytime so long as you are connected to the net). However, my conversion with them is not so high and here the price and time definitely kicks in. If I am near a store (Barnes and Nobles, Borders, or any other), many a times it is more efficient for me to just swing by and get what I want than pay extra (shipping) and wait a few days to get it. Now this situation is changed if I am giving a gift in which case I do end up buying it from Amazon.

    Bottom line, as you suggest yourself, I also think that even the best targeted marketing will not result in a proportional sales unless other criteria/needs are satisfied. For sure it can play a big role in the ‘acquisition’ arena but the site has to address the question of “what next” once the traffic arrives at their site.

    As to your comment about Amazon withholding access to book research — well, I personally think that would be a bad idea as there is no dearth of research avenues (at the minimum there is always Google 🙂 ). I think the stress should be more on offering a value-proposition to the consumer that they just cannot refuse — and this need not be the cheapest product but more likely an interactive vector product of priceXconvenienceXavailabilityXurgency (just as an example).

    Bottom line – analytics does not create needs (and by extension sales/conversions). What good analytics does is to arm the various folks in Sales/Marketing/Pricing with the knowledge necessary to exploit those needs by creating a good proposition.

    Enjoyed the read.

  3. Hey Ned! I can see you are as passionate about this as I am. As analytics friends we want to give analytics as much credit as possible. But I suppose the trouble with being analytical is that you analyze yourself too and only take the credit that you really can take 🙂

    I enjoyed a lot sharing this moment of thought together, thanks to your comment!

    Good points on shipment costs and swinging by B&N. You must live near a well stocked B&N though. I came to realize that most physical book stores carry only a very limited selection compared to online. Seems like book stores have become places of inspiration and entertainment today more than book search places. But thanks to “check availability in store” function, I get what you are saying.

    Have a good start into the week.

  4. Hey (Michael) graywolf!

    That is pretty amazing, I must say! Three cheers to 3G.

    Ned raises a good point with cutting down time even further, namely by avoiding shipment time.

    Unless you live in Manhattan maybe where Amazon was experimenting with same day delivery. But if you do live in Manhattan you probably come by B&N on the way to / from almost anywhere to anywhere. Well, not quite.


  5. Akin, you are right that all physical stores are not well-stocked. I used B&N as an example but the point I was trying to make was that I am lucky enough to live close to several physical stores (B&N being one, but there is Borders, Davis Kidds and a couple of others) and it is worth my time to do a check for the availability in one of those to save shipping $$ and time.

    And I got to agree with your statement about Bookstore for inspiration and entertainment :-). Seems like many are gravitating towards coffee-houses with books being the secondary reason for visit.

    You too have a good week.

  6. Just a side comment… my online shopping preference is all about finding the best deal and checking out the store’s reputation. If the store’s selection, cart and checkout processes aren’t Amazon-level, that’s fine as long as the store’s reputable and the price is the best price out there

  7. Marc,

    Totally agreed! I still remember ordering a cell phone from some NYC tech department store. Boy, was the return and customer service process a nightmare! Back those days everybody thought they could be an eTailer if only they had Fedex come by and pick up parcels. Not so, it takes some real customer service and professionalism.

    But isn’t the online reputation thing a good one, thanks to customer ratings and reviews. When you walk into a physical store you have no idea what other customers experienced. To that effect, Yelp is awesome too.


  8. Nice post Akin. As someone who was on’s payroll for a couple of years after they acquired Alexa Internet, I feel compelled to chime in.

    #1. There’re not always the cheapest. Other players like can undercut them. But Bezos’ sharp mind turned lemons into lemonade. They took a somewhat failing venture – “zShops” – and transformed it into a monster via the “New and Used From…” option.

    They made a rule for other vendors looking to tap into Amazon’s huge and loyal customer base: You can sell your wares here to our customers, but ONLY if you sell for the same or LOWER price than us. WHATT?? Yes. Extremely counter-intuitive, but solidly logical. Sure, is going to “lose” the sale to competitors selling to their customers on their site when the competitor has it for a lower price.

    So what would gain from this? A) It pretty much guarantees their customers base keeps coming to because there they’ll find the lowest price even if it’s not from directly (“one stop shop”) and B) still gets a cut of the sale, without the headache of logistics, warehousing, shipping, and carrying costs.

    That one move puts Bezos in the CEO Hall of Fame in my book.

    #2. Often when they are the cheapest, they’ve got damn solid math behind it. They sell at a loss on a number of SKUs. But it’s strategic. Same way Walmart built their empire with their “opening price point” gambit.

    #3. They also play the “Turkish rug merchant”. Given they’ve built up the most consummate repository of detailed, quality and authentic consumer product reviews, you go first to research. In effect, they’ve spent two full hours with you explaining all the ins-and-outs of how you can differentiate quality rugs from lesser quality rugs.

    And they’ve been serving you hot mint tea the whole time. Sure, you could wander around the dusty Web and potentially find it for a few bucks cheaper, but that just wouldn’t feel right. You’re warm and cozy, and you’re just one-click away from having the merchant slap an international mailing label on the rug and you can get on with your vacation. Why go anywhere else?

    They’ve been data-crazed from the beginning and actually put their money where their mouth is. I’ve got some upcoming posts on my site where I discuss determining a/b test length and prioritizing test concepts. Some of what I discuss corresponds with the philosophy and practice around marketing experimentation at

    • Thanks much Jared. And I see you must have gone through the mint tea and probably have the rugs at home to prove it.

      It is usually apple tea in the bazar in Istanbul. And you are right, once you drink it you will definitely and very gladly buy the rug, or the leather jacket.

      Very interesting analogy to Amazon’s rich “reviews as customer service.” It really is another beautiful example of marketing as a service.

      Agreed. Hats off to the market place idea too. What an unprecedented move that was.

Leave a Reply

Your email address will not be published. Required fields are marked *