Deprecated: Function set_magic_quotes_runtime() is deprecated in /home/mwexler/public_html/tp/textpattern/lib/txplib_db.php on line 14
The Net Takeaway: Page 24


Danny Flamberg's Blog
Danny has been marketing for a while, and his articles and work reflect great understanding of data driven marketing.

Eric Peterson the Demystifier
Eric gets metrics, analytics, interactive, and the real world. His advice is worth taking...

Geeking with Greg
Greg Linden created Amazon's recommendation system, so imagine what can write about...

Ned Batchelder's Blog
Ned just finds and writes interesting things. I don't know how he does it.

R at LoyaltyMatrix
Jim Porzak tells of his real-life use of R for marketing analysis.






Gift Cards: Turning money into air? · 02/07/2005 04:35 PM, Marketing

Many retailers have been experimenting with “Gift Cards” as a replacement for the old gift certificate. In addition, malls are offering cross-store cards, which is an interesting idea. Banks ahve also offered “debit cards” which are basically traveler’s checks in a card, usable anywhere Visa/Amex/MC/Disc/etc. are accepted. However, as usual, the stores are finding ways to treat what could become great customers… as garbage.

Now, before I get into this, please understand where the stores are coming from. Financially speaking, the money a consumer spends on the gift card is not bookable as direct revenue like it would be if the user bought a product. Instead, its like a half-transaction: the store has the money, but since the gift card is a liability until it’s redeemed, it sits on the books as “money owed”. So, like any debt, the stores want to amortize it over time and get rid of it if they can. They should be doing this by running promotions to stimulate usage… but instead, they take the “screw the consumer” approach.

This post is stimulated by a recent article in the Washington Post about yet another suit against Simon Malls (if you don’t have a UN/PW, try this Bugmenot search pre-set for the Post) because they administer various fees and expire their cards over time. Now, this is considered illegal in various states (including Mass., where I and many Simon malls live), but the stores are fighting these state laws tooth and nail.

I’ve never really understood the role of gift certificates and cards like this; my wife’s family gives the gift of ultimate flexibility: cash. And not checks either; they want it to be spendable that night. Gift Certs reflect this American approach that “giving cash is tacky”, which keeps stores happy. But in their zest to protect themselves, they go a bit overboard.

Here is some of the Post article:

Simon Property charges a $2.50 monthly administrative fee on the cards beginning six months after it is issued, according to its Web site.

Spitzer also says a $5 replacement fee isn’t disclosed on the company’s card as required by the law.

According to the Simon Property Group Web site, customers can prepay from $20 to $500, plus $5.95 shipping and handling charges, and give the card as a gift. The $2.50 monthly administrative fee begins in the seventh month the card is issued, until the balance is zero. A customer whose card expires with a positive balance is charged $7.50 for a replacement.

Now, as a marketer, scewing my customers seems wrong. I would like to propose that there is a strategy to facilitate both purchase of the cards, utilization of the cards, and happier, more loyal customers as part of the process.

There are 2 types of people who buy these types of products: Those that like the store and want to spread the message (Influencers) or those who know that the recipient likes the store, even if they themselves (the purchaser) doesn’t (Facilitators).

Influencers shouldn’t be screwed by giving a card that loses value over time; instead, you reward the Influencer with additional value on their card (of course they have one, they love your store) if the recipient uses the card earlier. Decline that bonus over time, and see how the reduced carrot works. This transitions nicely to a full rewards program if you don’t have one, and is an easy link if you do.

Facilitators are “bonus” money who are not usual customers… but they could possibly become one. If nothing else, they create additional spend in the recipient, given that all of that money could have gone elsewhere. We can still reward the facilitator by enhancing the value of the card. Instead of expiring the card, we make it part of a special program which gives the purchaser an additional discount if they buy 2 more cards during the year. This program would also be part of a larger “card-holders are great” program, which gives the recipient access to special card-holder offers, (of course they are delivered mostly by e-mail, but you knew that was coming). Finally, most stores are in the same mall with other stores the Facilitator might want to shop at for themselves. So, set up a swap of value between your retail chains so each side can get credit at the store they want, following the reward program above. In fact, you can even use this as a way of letting the recipient know where to buy a gift card for the gifter!

So, personally, I think these things are not such a great idea. But instead of fighting and finding ways to screw over the users who do like them, turn this process into a way to stimulate spending, and not only get the card value booked, but turn it into a way to generate repeat and loyal sales from both the Influencer and the Facilitator.


* * *


I Hate "Tags" · 02/04/2005 04:14 PM, MetaBlog

Update: Lots of traffic to this link from some very nice blogs and users… so, note that it rests in context with my complete series (so far) on why I dislike tagging, including:

I Hate Tags

I still hate tagging….

I continue to despise tagging…

In conference

Now, back to your regularly scheduled reading…


There is this whole movement currently about lazy ontologies, referred to as “tagging”. I started to mention this about my complaints on the Simpy bookmark manager site, in my post about Furl.

This has started to gain steam with the addition of tags to other information management sites (Sorry, I meant “social” sites, which are really the same thing, even if no one believes me). For those not familiar with the term, the idea is that people can just label information chunks (bookmarks, photos, blog entries, wikipedia entries, whatever) with whatever term they like, and if you are want to use that categorization to later locate this information or related concepts, you have to guess what crazy word other people are applying to the concept this week.

For example, on this site, I might label a posting as being in the “Database” category. Others might prefer to tag it as “Oracle”, “Data Mining”, or whatever they want. For their own personal use, that’s grand, but when its part of a shared social structure, its not really helping. And now the “bologsphere” has decided that its more useful and powerful than an ordered and structured ontology.

David of Technorati fame has written about it for the AlwaysOn network, read it here.

He mentions this article by Clay Shirky about this rise of unconstrained taxonomies, called folksonomy.

Now, both of these guys are very, very smart… and they are very, very wrong.

A few points from Clay, since he started it all:

“The advantage of folksonomies isn’t that they’re better than controlled vocabularies, it’s that they’re better than nothing, because controlled vocabularies are not extensible to the majority of cases where tagging is needed.”

Um, ok, so we can say that doing the wrong thing is better than nothing? And if a controlled vocabulary is not extensible, then its the wrong choice. A properly designed taxonomy is extensible by its very nature: It describes and organizes a set of data.

“Furthermore, users pollute controlled vocabularies, either because they misapply the words, or stretch them to uses the designers never imagined, or because the designers say “Oh, let’s throw in an ‘Other’ category, as a fail-safe” which then balloons so far out of control that most of what gets filed gets filed in the junk drawer.”

Umm… again, we say we don’t like taxonomies because they get misused? So, its the screwdriver’s fault that its hard to nail with it? Its so hypocritical: People are more than willing to accept some difficulties of working with tech, but then unilaterally decide “Oh, this is too hard. I don’t want to learn how I should to it so it works; I’ll just throw my way at it. After all, it works for me, why not make everyone else have to figure it out?”

Because that what it comes down to: either learn the proper use of a shared taxonomy, or try to figure out how each person chose to organize their content. If you are trying to share, I don’t think this latter is the best approach. It’s like telling people your address, but using non-euclidean coordinates. If they can figure it out, they are welcome to visit.

“Any comparison of the advantages of folksonomies vs. other, more rigorous forms of categorization that doesn’t consider the cost to create, maintain, use and enforce the added rigor will miss the actual factors affecting the spread of folksonomies. Where the internet is concerned, betting against ease of use, conceptual simplicity, and maximal user participation, has always been a bad idea.”

Basically, he’s transferring the cost away from the tagger… and onto the user. Actually, the best bet is simplicity x utility. Ease of use is not the same as utility: folksonomies will stay only as long as good terms are chosen that others can “grasp”. Once we recognize that we are all using 20 different terms for the same thing, and that’s making info hard to access… then we recognize (sigh, yet again) why an organized typology makes sense.

Ok, this is long enough. You get my point: The relaxing of standards around the information organization increases the cost to access the information. We rely on Google now, but we should have other ways to get to information and enhance it than phrase presense. “Folksonomies” are just delaying the useful organization in terms of a short sighted, fun approach.

And I like fun, but when you need to find something and can’t because people are using trendy and impossible to understand terms to organize it… suddenly, it ain’t so fun.

PS: Who am I to argue with the amazing Dan Bricklin, but I disagree with him too, to some extent: My paraphrase of his approach is Authors shouldn’t tag, but readers should. I think if readers would tag with an eye toward use by others, then maybe its ok… but we know no one will.

PPS: Good article in Salon unabashedly supporting tagging with some great quotes: “This isn’t a big technical innovation,” says Ross Mayfield, CEO of Socialtext. “It’s more the simplest thing that could possibly work, that shouldn’t work, but happens to.” and “The system doesn’t have to be perfect to work well enough for participants to find it useful”. Again, after using all the major “tagging” sites and approaches, I have yet to see much of an aspect of “works” and “useful”, unless you define them as “entertainment” and “fun”. Perhaps I am using the wrong tags to understand tagging… but then, that’s my point, isn’t it…

PPS: Danny Sullivan is, as usual, smarter than the average bear, and he sees what I’ve seen: tagging isn’t new, and hasn’t, to date, shown huge value for searching…

Comments? [6]

* * *


To the scumbag who keeps referrer spamming my site... · 02/04/2005 12:26 AM, Trivial

Don’t think we won’t destroy you. The funny part is, you won’t ever know quite how we did it.

For more info, feel free to follow any of these links:
Shetef Spam


* * *


Great Oracle stuff at DBAZine... · 01/22/2005 04:03 PM, Database

I always learn something new from these articles. Check out their most recent collection:


* * *


Yahoo Desktop Search Beta launches... · 01/11/2005 02:02 PM, Search

YDS is a repackaged X1, which I examine elsewhere and liked but not enough to pay for nor fully recommend; I repeat my concerns below.

A great review of YDS at PC Magazine points out the stuff not mentioned elsewhere:,1759,1749966,00.asp

“YDS has almost all the features of X1 Search version 5, which will be released by the end of January, along with a Search Web button that links directly to Yahoo! search online. The main X1 features not present in YDS are support for Eudora and Mozilla e-mail clients, the ability to index offline PST files in Outlook, and support for indexing remote drives across a network. While YDS and X1 are very similar at present, future versions are expected to diverge.”

The biggest annoyance to me is the lack of support for network drives, but that’s to be expected. After all, consumers never have more than one PC, right (sarcasm dripping)? As always, these attempts to limit “consumer” versions often mean users will choose products which DO support the features they need, even if they are “business type features” used in your very own non-business house. For example, the admirable Copernic Desktop Search supports network drives without an issue. Note that CDS has its own problems, which is that you have to manually tell it what filenames to store beyond the usual document files, a pain if you use any type of file extension besides the most common ones (and if you use any program besides Office, you probably do). That being said, its pretty good in almost every other respect.

If you want to try the Yahoo X1 offering, check out:

Note that in my examination of X1, it was very good, almost the best, but it phoned home too often for my liking given its excessive price. Since this one is free, a bit of phoning home may not matter… but I don’t know how much of the phoning home is turned off here. I just don’t like products which do it, notice or not, and if you can’t turn it off, then its a no-go for me.

X1 has taken this opportunity to release a “workgroup” edition which centralizes some management and storage issues. In the meantime, they continue to sell the desktop version for $75. That’s a steep price to pay to get no yahoo branding and network drive support.

Comments? [1]

* * *


Original Sandisk Cruzer... · 01/10/2005 03:19 PM, Tech

(Even More Updated: Sandisk changes the game with this clever approach.)
(Updated: Users have pointed me to a few other options:)

What a great gadget. I bought it mostly because I needed to buy a 256mb Secure Digital card, and it was only $10 more to get it with the Cruzer reader than without. Recently, needing to bring a bunch of presentations and other material to an off-site meeting, I realized this was the perfect solution, and pulled it out of the desk. Its been back in my bag ever since.

The Cruzer, model SDCZ1, is a USB mountable memory drive which lets you swap out SD cards. So, its either unfillable, or its a card reader, see it as you will. After seeing this, buying “hard-locked” or non-expandable USB drives seems silly, because whatever you buy, it’s going to fill up, and it’s non-upgradable.

This original Cruzer came with a little carry case and a short 3 inch USB extender. The plug slides out when the thumb notch moves forward, middle is “neutral”, and when slid to the other end, ejects the SD card.

Original Sandisk Cruzer

But there are some bad sides. It is about twice as wide, 1.5X as thick, and a bit longer than the modern USB drive, though not heavier. Also, its USB 1.1 only. Third, it’s discontinued.

The Cruzer name has gone on to become the name of the entire non-expandable line.

Good news is that Sandisk has made a revised version which is not as portable and not as clever, but close in size.

The ImageMate® USB 2.0 Reader/Writer doesn’t have the thumb switch for the slide-out plug, and it doesn’t have a case or a cute little cable. But the basic idea is there: you insert a card, it mounts, and its like a portable drive.

“Buy” the way, if you want to purchase this new one, as usual, appears to have the best price, though NewEgg, when they have them, is usually not far behind.

But I think Sandisk really missed the boat on this one. The Cruzer was about as clever a USB drive as I could want (do I need yet another mp3 player? How about a glow in the dark case? Please…) and I am really sorry that they were discontinued. The combination of clever design and pocket size made it easy to throw in a bag and never fear of obsolescence; this new one is large enough to become “yet another thing with a cable to carry around”… and so I won’t. Too bad; yet another product ahead of its time and now gone.

Since original publication, users have pointed me to other options:

Oh, and why, why are the USB ports right next to a bezel edge on so many machines? Who designs this junk?


* * *


Why are Men's Wallets so Huge? · 01/02/2005 09:12 PM, Personal

UPDATE: Lifehacker has a nice collection of posts about wallets small and large… Wallet Hacking Roundup

Ok back to the original article….


What a joke. I was going to buy a new wallet, since the old one had so many holes that my money was falling out.

But the new world is different than the old one. Because there are so many checks for ID, now every wallet has a flip-out compartment for driver’s license or other ID cards, aka the “pass-case” wallet. The wallets themselves are all larger, in part due to more credit card slots, but in part because (I think) people like bigger wallets to feel like they have more money to carry around.

My current wallet, with all my junk in it, is about 3/4” to 1”, depending on how hard I press it. It’s about 3” tall, and 4” from fold to edge. I really just want something that matches its slim dimensions, but the current models out there are all way too big.

Let’s look at some current wallets out there:

In fact, after all my hunting, I still can’t find a bi-fold wallet which is 4” or less W, and 3” or less H. “Card cases” which are mostly smaller, don’t hold money, just credit cards.

If anyone sees a brand of wallet designed for normal size pants pockets, please let me know.

Comments? [3]

* * *


Tuple-canoe and Tyler too... · 12/28/2004 06:50 PM, Database

First off, if you don’t get the terrible attempt at historical humour in the title, read the history of John Tyler, 10th president of the US.

When reading about databases, you may encounter various terms that seem like strange words for stuff you already know. Don’t waste your time looking them up; I’ve translated them for you here.

Cell=Field=Value=Data point (these are all intersections of Rows and Columns)

Yes, some people refer to the Columns as Fields, but to be honest, they’re wrong. Sorry ‘bout that.

Oh, and two other additional tips:
As seen in Speed Up SQL Server Apps by Roman Rehak, there are two common SQL queries which are poorly optimized in most database systems, and can be rewritten to help the database create a better execution plan.

1) Replace COUNT With EXISTS When Checking for Existence
A classic case is to code branches based on record existence, something like

     WHERE ShipVia = 3) > 0
  PRINT 'You cannot delete this shipper'

But if you are just looking for at least one (AtLO, or 1 or more, or however you like to phrase it), then an EXISTS clause is more effective. Basically, the database can short-circuit as soon as it finds the first row, instead of having to count them all (and throw away the value since you never really use it!).

     WHERE ShipVia = 3)
  PRINT 'You cannot delete this shipper'

EXISTS is also used in where, as in

  where exists (select 1 from t1
                      where t1.customerID=t2.customerID)

2) Be Careful When Using WHERE IN and WHERE NOT IN

If you are just doing WHERE customerID in (1, 2, 3), then things are probably fine. But if you have lots of values in the WHERE IN, then the database will often struggle. If you look at the explain plan, you may see many nested loops, which make the database work harder than it should. Instead, you might consider rewriting as an Outer Join… No, I don’t know how many “lots” is, so as Tom says, run it both ways and time it.

Replace the WHERE IN clause with OUTER JOIN if you’re using a subquery to generate a potentially large list. Doing so can improve performance significantly:

  SELECT c.*
     FROM Customers c
     LEFT OUTER JOIN Orders o
     ON o.CustomerID = c.CustomerID
     WHERE o.CustomerID IS NULL

The LEFT OUTER JOIN selects all rows from the Customer table—whether or not a customer placed any orders—and joins them with the Orders table. Then the WHERE clause filters out the rows where the columns from the Orders table have NULL values. Similar to the EXISTS thing previously, the database can “short-circuit” from having to do every possible join and will just deliver the necessary rows.

There are other tips in this article, many SQL-Server specific, but I thought these two deserved special mention as being very handy.


* * *


New Section on Net Takeaway · 12/21/2004 12:07 PM, Trivial

In case you hadn’t noticed, I have added a new section to the sidebar (one of the problems with RSS feeds: you never see navigation changes on the home site unless you visit. Perhaps its time to give excerpts instead of full content?).

Besides the ability to drill down on specific content categories (Analytic, Marketing, Tech, all in the drop down) and run a text search which have always been present, I have also made links to a few “special sections”. For example, Judoscript has a couple of entries about how to use it with tips, tricks, and sample code, so it gets a special area. These special areas do not show up in the RSS feed, and also do not contain any articles which have been published in the main blog areas. Instead, these are more tutorial and reference sections, and tend to be longer. Also, they do get picked up by the search engine.

I’ve now added an R area. For too long, there hasn’t been a document aimed at the SPSS user, the kinda savvy user who may want to play with R. Consider it a “Cliff’s Notes” version to help deal with some gotchas. So, I am taking my shot. Click on the R Statistical System link on the side to visit.

Assume that at some point, I may collect similar reference material for SPSS and Clementine, but til then, try the search engine and the categories to see what I’ve said about them (and I’ve said lots about them.

And though the company I work for mails for the NFL, the NHL, and the WWE, I am not a sports fan. But many of you are, so perhaps you might enjoy this article about using R to examine baseball stats… Analyzing Baseball Stats with R by Joseph Adler.

I’ll probably make another entry in the “special” area about various tutorials, but if you are interested in yet another quick one, try IBM’s Statistical programming with R which has 3 parts; only parts 1 and 2 are online currently.


* * *


Unofficial SAS Weblog · 12/18/2004 02:38 PM, Analysis

Jasan Calacanis and his merry band have added a SAS weblog… not sure how many people will love it, but can an SPSS log be far behind?


* * *


On a previous episode...

powered by Textpattern 4.0.4 (r1956)