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 10

OTHER PLACES OF INTEREST

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.

 

HOW DID YOU GET HERE?

https:
https:
https:
https:
https:
https:
nettakeaway.com
https:
nettakeaway.com
nettakeaway.com

 

 

 

Tag Clouds display problems with tags... · 11/07/2007 01:41 PM, MetaBlog

Tell me, does this make tagging any clearer?

Tag Clouds Gallery: Examples And Good Practices

To use one quote, which is completely correct but overlooked by most tagging fans, “Your Tags Are Not My Tags”.

You see, if we allowed the tag collection to truly evolve into a functionally useful offering for knowledge discovery, then we would have a collection of “shared” language terms which allow linking, and of course the individual esoterica allowing one to snarkily comment on an aspect of the content without really adding much value to the collective.

Instead, people tend to focus on the latter. They customize their tags, avoiding using words that others may like, as they are not “individual” enough. That’s great for you. It stinks for the collective. It creates a myriad of walled gardens, the very things I think tagging should be working to eliminate.

Click here if you are interested in my other rants against tagging (some of the most popular links on this blog).

I know, tagging is no longer supposed to do everything and the kitchen sink, thank goodness, but I also think people continue to assume that unstructured tagging, by itself, adds value. It’s part of the story, but without some consistency, its just a mess in most cases. I look forward to us continuing to clean up that mess.

Comments? [3]

* * *

 

We will control the horizontal. We will control the vertical. · 11/05/2007 01:11 AM, Marketing Tech

(From The Outer Limits)

All the fuss over Comcast restricting traffic on its network is bringing around (AGAIN) the fuss over whether an ISP subscriber is allowed to do whatever they want on the network. From “unlimited” bandwidth really being capped to the no-longer-true expectation that an ISP would include alternative data sources like Usenet to the latest fuss over Comcast restricting Bittorrent shares, people are finally starting to recognize that their ISP controls their view of the internet, and that the ISPs act like they have the right (via “license agreements”, click-wrapped or otherwise) to fog, reduce, obscure, track, and share anything that you do on their portion of the network (meaning, actually, anything you do, period).

And here’s one most people don’t think of, but should be included in the list of bad behaviors. I recently typed in a bad domain name, and I wound up on this page:

http://wwwz.websearch.verizon.net/search?qo=www.blahdeblah.com&rn=RmNKJQ-iZ6uE_dv\\

This page is basically an Infosearch powered tool for Verizon to make money off of search ads. If I click on the links, I generate cash for Verizon. Instead of “helping” me, its just a moneymaker. How do I go back to just getting an error, which is what I should be getting for a non-existent site (domain exists but has no site, so returns a DNS error)?

Oh, don’t worry, you’ll never find it. You have to go to “About This Page”, the button on the top right (which is an image, not even text… why is that?).

Depending on the Verizon page you get to, you either get something like this which does give a direct link to opt out (but wait til you see how to do it…) or you get this page, which tries to hide it even more, this page on Infospace, by putting it in a fold so you STILL don’t get the info until you click more.

And how, pray tell, do you opt out? Oh, simple as pie. You have to go into your modem/router settings, and change the hard coded DNS 4th octet from .12 to .14. What a joke. And catch the name, “DNS Assist”. In just what way are you assisting the DNS process by intercepting it?

Opting out of DNS Assistance

So, you force this moneymaker deal on the user, and you make them change their DNS settings in the depths of the network config of their router to opt out?

So, add another checkmark to the fight between the network and the user. I know other ISPs do this stuff, and practically every hotel network also does it. And of course, let’s not forget the immense success of the American mobile market, which involves carriers controlling every aspect of the user experience on their network, and charging for every point of control or access that they can.

I believe I am paying for the right to access the Internet, so get the heck out of my way. They believe that I am just giving them money so that they have the right to control my “onramp to the information highway” or whatever metaphor their marketers are telling each other when they enact these plans.

The time is coming when these networks will lose the little loyalty of the customers they have left. Games like the ones I mention at the top of the post are nails in the coffin. It doesn’t take an expert marketer to look at the “growth” of mobile, the impact of the iPhone, and the rise of consumer control of marketing to recognize that the money is in enablement, not dictatorships.

Or, given the behaviors of Comcast and Verizon, maybe it does.

Also mentioned at ConsumerAffaris.com Verizon Overrides Internet Searches With Its Own Results and GigaOm Verizon Redirects Typo Traffic to its Own Search Service

PS: Yes, it does not escape me that Yahoo!, my current employer, is providing the search results and ads on that “Assist” page. I am pleased for them to make some money, but I don’t think this is the way I want to get it. And yes, “my opinions clearly do not reflect those of my current employer, etc.”

PPS: BTW, if you really don’t like these kinds of games, there are Public DNS Servers . Note that a commonly suggested alternative to IPS DNS, the OpenDSN, also has a “OpenDNS Guide” feature for domain errors, and it cannot be turned off.

Comments?

* * *

 

Just how is this a reward? · 10/13/2007 01:07 PM, Marketing

USAir recently sent many of its USAir Dividend Miles users an email letting us know about their new policy. Here is the text:

Important Notice

Don’t lose your Dividend Miles

Dividend Miles #: xxxxxxx
Mileage balance: xx,xxx
Your last activity date: 06-13-2006

Dear MICHAEL WEXLER:

US Airways introduced a policy last year that rewards our customers for staying active in the Dividend Miles program. In order to keep your account active, you must earn or redeem miles within a consecutive 18-month period.

Our records indicate that you have been inactive since 06-13-2006. We want to make sure you keep the miles you’ve earned. To keep your account active and hang on to your miles, you have several options:

* Contact Dividend Miles at 800-428-4322 and pay a $25 preservation fee with your credit card. * Earn miles by flying on US Airways or any of our airline partners. * Sign up and earn miles with one of our credit or debit cards. * Use any of our other partners for everyday activities such as dining out, sending flowers and more. * Redeem your miles. * Shop with over 100 premium retailers for name-brand merchandise at the Dividend Miles Shopping Mall, where you can reactivate your account for as little as 99 cents.

Take advantage of any of these options by 12-31-2007 and your account will remain active for another 18 months.

OK.

To be clear, if you were rewarding me for staying active, then you would be giving me bonus miles, more than the usual, for being active (say we defined active as 1 flight every 3 months). What you’ve described here is actually a punishment for staying inactive.

And these options aren’t all that palatable. Almost all of them involve cash. If I were to actually use some of the miles, that wouldn’t actually change the expiration date for the rest. And that “99 cents” they mention? Couldn’t find what any merchant was offering for 99 cents. Yep, this just feels more rewarding every second.

So, note to copywriters: Air travel already sucks. Trying to make yet another punishment look like a reward is not wise. Note to marketing managers: reward programs need more than just give and take of rewards; you need to make it part of the experience. I have many miles on your books, but I don’t fly you any more. Is this the way to win me back?

Lots more commentary at the Consumerist:US Airways Rewards Customers By Threatening To Charge Them $25 For Inactivity.

Comments? [2]

* * *

 

Adding percentages? You're wrong. · 09/28/2007 03:03 PM, Analysis Marketing

A really good paper on something most stats folks know but most everyone else forgets: Percentages applied in order are not themselves additive. You have to “do the math” since the base changes after each percentage increase or decrease.

So, the classic case: You have a 20% off coupon, and the item is marked off 25% from regular price. What’s your actual savings at the register? Most people immediately add 20% and 25% and arrive at 45% off, immediately running to the register.

But if the price was $100, your final price will be $60. That’s only 40% off. You thought it was going to be a better deal, didn’t you?

The paper mentions an even more insidious error in its intro. “A 60% decrease followed by a 70% increase on the standardized test scores in the state of California seemed to cheer up a lot of people”, but they follow with “though this actually results in a net decrease of 32%”.

Percentages are very ambiguous things. Sometimes, it makes sense to explain an event with an understanding of its context, and that’s something that percentages can help with (10,000 may be a lot, but not when its out of 10,000,000,000… 0.0001%). However, big and small bases can cause vast errors of judgment (if I get a 66% click rate on a message, that’s amazing. Its less amazing if its 2 people out of 3 mailed, or 2 clicks on 3 impressions.) I always advocate setting the stage by warning people if bases are small or large.

And now we have a reminder that combining percentages is more devious than you expect. Don’t be fooled! Do the math, or do the time. (Ok, that last wasn’t quite warranted).

BTW, if you really, really want to quickly add the percentages in the above example: 25% off is the same as multiplying by 0.75. 20% off is the same as multiplying by .80. When expressed this way, you can multiply (NOT add) them together. So, .75 * .80 =0.6 (the % of the original price remaining); 1-0.6 is the discount (0.40, or 40%). If you are augmenting (adding 25%), then just do 1.25. I know, this is simple math, but as the paper points out, most people forget what they learned in 9th grade so many years ago.

When Two and Two is Not Equal to Four: Errors in Processing Multiple Percentage Changes by Haipeng (Allan) Chen and Akshay R. Rao (link to PDF)

(Thanks to Dr. Dobb’s editor Jonathan Erickson for mentioning this paper.)

PS: Ok, for extra credit:
100 + 20% + 25% + 60% = 240
Now, remember what we said? Each percentage has to be calced in order since the base changes each time?

So, what do you expect from this, reversed order, with a large base as the first calculation:
Is 100 + 60% + 25% + 20% bigger or smaller than the first one?

Right, trick question. I didn’t say if the sparrow was African or European. And yes, both calcs wind up with the same answer.

Comments? [2]

* * *

 

Ubuntu 7.10 and No Mouse on Virtual PC · 09/28/2007 12:30 AM, Tech

I hope I don’t have to keep writing this same post over and over again. I just tested Ubuntu 7.10 beta, and like 7.04, the mouse is not recognized in Microsoft Virtual PC.

As I said in Ubuntu 7.04 and No Mouse on Virtual PC, the kernel has a bug which means that the mouse is not recognized in Virtual PC. In addition, the same graphics bug is present (when you get past the [OK] text stuff, you get a scrambled screen) but if you choose the “safe graphics” mode while installing from the LiveCD, you can at least test it in comfort… once you get the mouse working.

So, just like I said there, the workaround is simple, but annoying.

From https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/87262 which points to https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=223606, where both say to include the kernel parameter i8042.noloop on the boot. This can be done manually each time on the liveCD (Press “f6” for additional options at the boot screen, and hand type it in on that text line) or if you install it, you can manually add this to the boot config file. On my tests with VirtualPC, this allowed the mouse to work as expected!

http://grumpymole.blogspot.com/2007/05/ubuntu-how-to-edit-grub-boot-parameters.html tells how to edit the boot config file.

Yes, it is very annoying that after all this time, neither bug has been fixed. I’ll log them again, and we’ll see what happens.

They don’t call it “Open Sores” for nothing.

PS: That graphics issue? Its because the default bit depth (colors) for Ubuntu is 24-bit, which Virtual PC does not support. Yes, Ubuntu could check to see if the graphics card supported 24-bit color before tossing the user into this, but since most real hardware supports 24-bit color, its another “mostly affects MS VirtualPC people” issue. Anyway, if you install Ubuntu to a drive, you can just configure the colors properly inside Ubuntu and the boot will go smoothly after that. Basically:
1) Start a command prompt window (Applications, Accessories, Terminal for most folks)
2) cd /etc/X11
3) pico xorg.conf (pico is the easy editor most linux people don’t talk about; nano is another similar editor)
3a) If you get a complaint, try running sudo pico xorg.conf, which is a way of saying “really, let me do this”, so to speak.
4) In the “screen” section, change the defaultDepth from 24 to 16.
5) CTRL-X to exit (saving when prompted of course)
6) Either type reboot, or just use your favorite way to exit

If you love the LiveCD approach, then keep “Starting in Safe Graphics mode”

PPS: Another nice summary with a step by step is at Virtually Vista by Mike Kolitz of Microsoft.

PPPS: Finally, the Ubuntu wiki pulls together a bunch of helpful info around Ubuntu and VPC: https://help.ubuntu.com/community/HowToConfigureUbuntuForMicrosoftVirtualPC2004

Comments? [1]

* * *

 

Make It Viral? · 09/25/2007 12:07 PM, Marketing

Folks keep asking me how they can “make things viral”. That’s a somewhat silly question, but advertisers still believe that they can control people’s action a bit too much.

Here’s a somewhat different way to look at all this viral stuff. Remember, way back, when you learned about conditioning in your PSYC 101 class? There were 2 kinds: Classical and Operant.

Classical conditioning involved presenting a stimulus and trying to get a response. You “rewarded” or “reinforced” the correct response (or behavior), with the goal being to get an association between the stimulus and the response. When that association was in place, you could show the stimulus, and the response would happen automatically, even without the reward. You’ve been waiting for me to mention it, so yes, this was the one with Pavlov’s drooling dog.

Operant conditioning focuses more on the reward part. This model doesn’t actually need a stimulus; it assumes that the behavior will exhibit on its own, and you reward the behavior when you see it. But you can vary the pattern, size and type of rewards to create faster or more long lasting learning. This was the B. F. Skinner work, and its worth reading about his many contributions to the world, from attempts to create world peace to the “Air Crib” at Wikipedia’s B. F. Skinner entry.

So compare these: One involves creating a stimulus and trying to get people to take action on it. This might be akin to creating a video, pushing it everyone’s face, and expecting people to forward it. This, btw, is how most people try to make things viral: They spread it from left to right, manually, and hope somewhere, somehow, it sticks. But how do you reward the person who does choose to send it? How do you associate your stimuli with the desired action (“Wow! An advertiser created video evoking hip irony in a self-effacing way! I must forward it at once!”)? Well, actually, its pretty difficult, hence the many failures. If you want to get technical, there are all sorts of issues such as stimuli recognition, reward association, the need for repetition, fatigue, etc. But you don’t need all that research to guess that this type of manipulation has lots of failure points, the most obvious being the authenticity, the clear “manipulation” of the whole thing. And once its spread all over, its hard to create the proper reward experience (which I describe below).

Ok, now look at the operant variant: You create what you are going to create, and if people do forward it, you reward them somehow. But Wait! Don’t give them points, money, or other aspects of material goods. That bribe will convert a potential reward into a punishment (no, not a negative reinforcement; that’s when you stop the pain to create pleasure, not create more pain). Why? Again, its relevance and authenticity (and cognitive dissonance): if you are rewarding me for doing this, then I’m not doing it for myself, I’m getting paid. And that’s not cool.

No, the trick is to evoke a sense of peer and internal reward: The sense of being the “supplier of that which is cool”. Remember, you don’t expect everyone to be the “early adopter”... but everyone has a group for which they are the “cool finder”, that mailing list they send cool stuff. You want to provide content that allows someone to experience the “I sent the cool thing before my friends had heard about it”.

That internal reinforcement will do more long lasting good than any material bribe for sharing. The initial joy of “this one’s pretty cool” is compounded by the “I’m the spreader!”. The work, therefore, is in making something actually cool (hint: break all your “rules” and don’t worry about what might work, consider it all an investment in learning).

Now, you might ask, how is this helping my brand? Well, small group cool finders (aka influencers, etc.) are filters, who spread authenticity and selection criteria. If they choose to share, then your brand is imbued with their halo. Again, no direct reward necessary or desired.

(To be clear, I am differentiating from rewarding small core dissemination groups. I am all for giving goodies out to my beta testers, or to my early fans, or to those who sacrifice their time out of love of my brand/product/band/service/whatever. But keep it relevant, authentic, low ebay-resellability, and you’ve got the right model).

So, make it viral? Fine. You make it so cool that people want to send it, and it’ll make itself viral. The marketer/agency/media site’s job is to make sure that it’s available, searchable, findable, leakable, and shareable. Viral things are like drugs in that they can be addicting for a while (Digg sucks up way too much of my day)... but unlike drugs, when you “push” too hard, you actually lose customers.

(Yes, that last simile sucked. I’ll find a better one later).

Comments?

* * *

 

Web Sites are Journeys... · 09/17/2007 01:15 PM, Analysis

The aspect of narrative, of understanding user clicks over time, is a hugely missing part of most web site analytic tools and approaches. (Its also a missing part of most marketing campaigns, but I digress.)

Here are some good links to get people thinking about what the next generation needs to look like, by looking at this approach to designing sites:

Web Sites Are Journeys Not Structures

An Introduction To User Journeys

Comments? [1]

* * *

 

J vs. R · 09/14/2007 12:15 AM, Analysis

When we think of “statistical” programming languages, most of us in the stats field think of S and its open source cousin, R. I have a ton of content and information on R, reachable from the menu in the right hand side labeled R Statistical System. And so, having worked with SPSS, SAS, Stata, and R/S, I thought I had seen it all.

But I was surprised to see mention of a clustering algorithm implemented in “J601”. Some searching revealed that J is the next generation of Ken Iverson’s “APL” programming language. APL was known as the programming language with all the strange characters: comp sci people loved it, and everyone thought it was a nightmare. See some samples at the Wikipedia page on APL.

So, J appears to be a combination of the terse syntax of APL with a “functional programming” approach. Like R, it has a large collection of built in statistical and analytic functions optimized for matrices of data, and is appears to be used mostly by quantitative trading analyzers. While most of the stats/data miners I know in the marketing field tend towards the languages I mentioned earlier, its always good to know about the ways people solve problems in other fields.

See more about “J” at its home page, JSoftware.com. After looking it over, I think R is much easier to understand, but who knows? This could be the language you’ve been looking for…

BTW, if you are interested, another popular quantitative trading analysis language also derived from APL is K, which is part of the Kx column-oriented database system (which merges in-memory with disk in some impressive ways). See more at Wikipedia’s K page and it’s home at Kx Systems.

Comments?

* * *

 

"Backlinks" · 09/12/2007 06:19 PM, MetaBlog

A relatively recent change in some leading blogs is leading to lots of teeth knashing and annoyances. Its not just a UI issue, or a journalism issue, or an SEO issue… its also about greed.

The latest post on this issue is from Louis Gray but you can see it on any recent post at TechCrunch.com or Engagdet.com.

Basically, almost half the links in their entries link back to category or tag pages in their site, instead of the obvious target. So, for example, the links above to Engadget would actually not take you from my site to www.engadget.com, but instead to another page on this site listing other entries I’ve written about Engadget. If you click on a link which looks like an external site, you naturally expect to go there. And if its ambiguous, say a story about a person or a place, the hotlink text should say where it’s linking you, either through a symbol or in the hypertext itself.

Ryan Block, the editor of Engadget, has posted his explanation of why he does this but it reads as you’d expect: a dancing around the issue and a defense of a practice which is reflects a pure greed: keeping people on your site when they don’t expect to go creates new pageviews, and therefore ad impressions. If you want to help users, you make sure your links clearly communicate where they are going.

I enjoy reading Techcrunch and Engagdget, and its a quandry. Because I hate supporting sites which do scummy things. Let’s hope the noise get’s them to change their practices. If you are writing about events and items which are not part of your enterprise or organization, and you include a link, readers expect such a link to direct them to another source, or even the actual home of the event or item. If you want to direct people inwards, simply use the language and give users a clue that you are sending them to the archive. Its easy to do… and because its so easy, the choice not to do so is very telling.

Look, there are lots of scummy things out there (did you see the free page counters who hide 3rd party links in your pages to get free search ranking increases?) and this is not worth hurting people over. We just all expect better from the leading bloggers (“they’re just like us, only they make money from this stuff!”) so when they abuse their position or our trust, it feels worse than if some desperate newspaper does similar tricks.

PS: Even the big guys dislike it: Look at Jeremy Wagstaff’s post entitled Sleazy Practices.

Comments?

* * *

 

SAS on Identifying and Overcoming Common Data Mining Mistakes · 09/08/2007 01:41 AM, Analysis

A very nice paper on data prep, understanding, and mining. Helpful not only for beginners but also those from other analytic disciplines who aren’t used to the need to prep data specific ways for different analytic approaches. A good read.

Identifying and Overcoming Common Data Mining Mistakes by Doug Wielenga, SAS Institute Inc.

Found via the DSanalytics.com blog which is also fun to poke around in.

Comments?

* * *

 

On a previous episode...

Admin
powered by Textpattern 4.0.4 (r1956)