Wednesday, December 09, 2020

Feedburner. Cool Name. What to do?

 I have some hobby blogs (like this one) that I have kept up on Blogger since around 1983. Or it feels that long. One one of them - (also known as ) - there is a feedburner widget which supposedly sends out my blog posts to my subscribers. 

I'm not sure it's working any more. I clicked around and found this:

 Uhg, there are an indeterminant number of people who were in that system. Feedburner seems impossible to login to. What to do now? Just remove it? Start again with what?

I just tested the signup feature to see if I could use it to subscribe. It seems that I can.
But how would I login and see my subscribers, the number if not the actual names?

I just realized that this blog also has a feedburner widget. It says it is owned by a corporate account that I still have access to. I'll try to access it.

Feedburner does seem active in that I got an email confirmation email that clicked thru properly.

BUT, the feedburner account in use relates to an old Google account and while I'd like to switch it to the new account that manages the current blogger set-up, I cannot begin to figure out where and how I would do such a thing... Help?

Wednesday, October 21, 2020

User Experience Design: Card Sorting

 Today I learned a new user experience (UIX) design technique. It's for organizing different topics into a few top level menus.  In the simplest form, there's closed card sorting. 

It starts by creating a card for each topic that the site is going to cover.  

With closed card sorting, a set of cards are given to different potential users along with a few pieces of papers with a topic on them, and the users are asked to sort the cards onto the pages based on the topic where it best fits. This gives guidance to site designers as to where to put topics in terms of where the users expect to find them.

Surprise! The Visitors Think About
Topics Differently Than the Professionals.
So Whose View to Use?

Open Card Sorting. To get more pure user feedback on how they visualize topics and categories, the cards can be given to users who put them in piles based on the categories that they imagine they should be organized along. This can reveal a more genuine sense of the mental maps with which users approach the relevant topics. It can also be overwhelming to users and in many cases, produces haphazard sets of logic that users turn to when they get frustrated and just want the exercise to end.  

A purer sort might be to give the users a blank set of cards and a few pieces of paper, tell them what the site is about, and ask them to put a major topic on each of a few pieces of paper, and then name and organize the cards. This system does not have a name that I am aware of.

It's easy for these techniques to get out of hand. It's important to remember that they are techniques to reveal the mental maps that people approach a topic with.  But the site designer, through careful wording, should be creating navigation and topics that steers users down paths that support the goal of the website.  Websites are not libraries or wikis where users are expected to freely browse and learn. Websites usually have a purpose and while knowing the mental maps with which users might first approach the site is useful, it does not necessarily dictate how the site should present its experience. 

Stay tuned or an example which illustrates these choices...

Friday, October 02, 2020

Patent to Build Reading Skills by hearing and seeing the sounds

Even as we rolled out SpellingCity, teachers and literacy coaches asked us to do more with helping students with sounds. They asked us to convert the games to focus not just on spelling practice but on practice activities for recognizing and working with sounds.  They wanted help not just with the spelling of words but with learning phonics and building phonological skills.  So we focused  on building the tools needed for games to help students with the sounds and the letter combinations that represent them.  The goal was to give students audio visual practice with the sounds that create words helping them connect the sounds that they hear and the letter combinations that they see.  

The idea was simple: We wanted to treat words like “tooth” as three blocks of letters which correspond with the three sounds: T, OO, and TH. But, as we searched, we could NOT find a system which mapped the sounds in words to the way the words are spelled. At first, this seemed unbelievable. Surely, in some university or research center, somebody had created a mapping which connected all the common English words into their sounds and mapped those sounds to the letters used to spell the words.

We spoke to a lot of people which  confirmed our initial findings. This mapping did not exist. Dictionaries, for instance, routinely have a phonetic spelling of words using various systems for writing phonics. But none of the dictionaries mapped the sounds back to the actual spelling of the words. Nobody had ever done this. Our vision came from watching endless tutors, teachers, and parents help students by pointing at a few letters in a word and having the student say the sounds that those letters created. We watched teachers help students read the sounds to decode the word and then blend them together to write them.

So, we decided to create the VocabularySpellingCity Phonics system, a novel contribution to literacy. The phonics system can be used for building a variety of prereading phonics-related skills including phonological skills, phonemic awareness, and spelling skills. Since we knew we had created something original and valuable, we started talking to lawyers. We decided in 2015 to file for a patent on our original system.  We started with two provisional patent filings. Our permanent patent is number 10,387,543, issued on August 20th, 2019. It’s called a “Phoneme-to-Graphemes Mapping Patent”. It’s a utility patent covering our original method for algorithmically mapping the sounds in English words to the letters. The patent grant is both a recognition of novelty, a recognition of usefulness, and a grant of intellectual property ownership. What is Phoneme to Grapheme Mapping? Phonemes are the basic sounds of the English language.  Examples of phonemes from the word “cheek”, would be: CH, EE, K.

 Graphemes are the use of letters to express these sounds.  In English, here are three example of patterns of how sounds (phonemes) are expressed by letters (graphemes): 

  1. Some sounds are created by a single letter which almost always makes the exact same sound. For example, the T is “ten”.  T almost always sounds the same (except when it’s in a combination with another letter like H). 
  2. Some sounds such as the long E sound can be spelled a number of ways including a "ee", or "ea", or an E followed by a consonant followed by an E at the end of a word, a y at the end of the word, and an "ey" at the end of the word.
  3. Some letters, like the S, can make different sounds. S usually sounds one way, like in sound, and sometimes sounds quite different, like in sugar (where it makes the SH sound)
 So how can this technology help?

 Students can hear and see the sounds by mousing over the sounds in each box of VocabularySpellingCity’s Interactive Phonics Boxes. Many classrooms have students first work on recognizing the initial sounds where the Sounds Boxes are used with images to match initial sounds. 
For commercial purposes, the patent belongs to VocabularySpellingCity. Patent 10,387,543 Holders of Patent 10,387,543 (current employees) 
The patent holders who are current VocabularySpellingCity employees are John Edelson, Obiora Obinyeluaku. and Kris Craig. The two xemployees are Jose Perez-Diaz and Harold Milenkovic.


Activities with Interactive Sound Boxes (that use this technology): Sound It Out,  Initial Sound SpellerFinal Sound Speller,  FlashCardsWord Study (available for logged-in students) and TeachMe More.

Sound-Based Activities for Phonological  and Phonics Skill Development:  Which Initial Sound?, Which Final Sound?, Initial Sound SpellerFinal Sound Speller,  SillyBullsSound It OutFlashCardsWord Studyand TeachMe More




Article 1

Friday, September 11, 2020

Link Beggers...

 It's amazing to me that link beggars are still out there begging...

I get a few of these every week despite all the spam tools that should screen them out.


From: ****<****> 

Sent: Friday, September 11, 2020 1:44 PM
To: T
Subject: Guest Post Request on :https://www.*****.com

  Hello dear
Sir/ Madam

I Need Guest post at Your  Site: 
With do-follow back link Permanent
Post and Insert link let me Know how much
price for each post
Waiting your good Reply

Sometimes I answer saying that a million dollars in bitcoins would do it.

Since I work for a legit corporation, this bit of humor will probably get me in trouble some day. Still, I think it's funny.

I was in an SEO meeting this week where we discussed link building and a suggestion was made to start asking for links. I was in disbelief but it turns out, they were talking primarily about contacting people who had written articles about us or were mentioning us. In these cases, you can sometimes turn the mentions into links, even sculpt them a little. And legit press will actually treat that as a profit opportunity and sell those links.


Tuesday, August 25, 2020

What's with all the links?

 To manage the process of optimizing our sites for the search engines (SEO), we license a tool called Conductor which helps us with tracking, ranking, and all sorts of the nitty gritty of SEO.

Mostly, the relationship is pretty minimal. We send them money, they let us use their tool (which is great).

This week we heard from them. Apparently, an unusual and alarming number of links were appearing aimed at our site and they worried that we were spamming or being made to look like we were spamming and were we aware of it?

Our answer:

"It is back-to-school season during a major pandemic and many parents are trying to figure out what to do for their children's education. Homeschooling is a school choice option that is now of interest not just to the ~1 million families that were doing it at this time last year, but all ~30 million families of kids with school age kids. The press as well as the public have picked up on this. Time4Learning is the leading homeschool online service. Yes. WE ARE TRENDING"

Nothing nefarious or devious about it. We're just getting a great deal of attention.

Thanks for checking. Oh, you too have children and have some questions? Well, check our website and look at the demos and videos, visit our facebook groups and ask some questions, download our free guide to starting to homeschool, and good luck to you."

Here's today's joker. It commemorates the 1934 World's Fair. It's part of the Americana collection of jokers. 

Monday, July 06, 2020

The Business Case for Podcasting

Lets assume that you are a business who runs an information site that makes its revenue on selling advertising to a specialized niche of people interested in that information.

For our purposes, the info could be anything. It could be about poodles. It could be about coffee mugs. It could be about business education for dentists. Let's assume, for the sake of this discussion, that the site is about poodles.

So assume there is an info website in place with people who visit it, people who write for it, people who sell ads on it, and people who buy ads on it.

Not Poodles but poodles is easier to spell!

As a growth and brand strategy, the info site would like to grow into all the places that people might want this info since people like to grow. But that is a lot of places. In some places, it might not make commercial sense. Overall, it would make more sense if the content could be created once, used everywhere. It would make less sense if the articles, videos, social media posts, podcasts, and so on were each an independent creation effort so there is no leverage and synergy.

In addition to a website full of articles info about poodles, the business might:
- syndicate its best content onto other platforms related to dogs, pets, families, and parenting.
- solicit other content writers about poodles and get their content onto the poodle website
- create videos about poodles and put them on Youtube
- create emails and newsletters about poodles
- hold conferences about poodles
- create social media accounts about poodles and fill them full of poodle info. This could include Facebook, Pinterest, Instragram, Twitter, and more.
- create ebooks about poodles and put them on kindle, iTunes, Google Play, and other channels
- create podcasts about poodles
- create new content on the new platforms as they emerge such as tiktok and interactive speakers (Alexa) and others that have not yet emerged

Is more always better? No, there is an optimal amount of expansion and content creation. You can  overspend on content creation.  You can over extend and put too much effort into too many media. You can create content around topics where there is no real advertising or business opportunity. Novels, for instance, are a powerful media but with NO opportunity for advertising. The color gray has a lot that can be said about it but virtually none of it attracts any advertising.   You can accumulate an audience but not monetize it.

What does, given all these possibilities, a business case for a podcast look like?

  1. Envisage success. If all goes well, in 12 months:
    1. What is the size of the audience that is following the podcast?
    2. What is the value, in terms of advertising, of this audience?
    3. These are easily answerable questions since we are buyers of ads on these platforms and we know what the revenue opportunity looks like. To think about it, a monthly advertising revenue estimate should be made.
    4. In 12 months, to maintain this audience, what are the monthly costs of producing and publishing this podcast?
  2. Measure progress.  Find some comparable successful podcasts and look at their history.
    1. How big was their audience after the first quarter, the second quarter, etc.
    2. Compare that growth pattern with the growth pattern that our podcast is showing.
    3. Are we going slower or faster than our model of success?
    4. Any adjustments we can make?
    5. What is the current expense to maintain the current growth rate?
  3. Create a spreadsheet and add it all up.  Does the business case make sense?  Is it a high ROI?  Is it a money pit?  
    1. What's the alternative? Could we for a certain amount of money sponsor some other podcast creator?
    2. Could the postcasts be published at the same time as videos on youtube and elsewhere?

Thursday, July 02, 2020

Hotwire Fision TV: Worst User Interface Ever?

I believe that when professors put together courses on designing user interfaces, the current Fision Hotwire remote interface should be used as an example of how awful an interface can be.

They obviously spent a lot of money on their slick remote gadget as it is thin and has a huge light up high res screen. But it is a nightmare to use each and every day.  It's so bad that you can spend hours arguing about what is the worst aspect of it and in the meantime, figure out that things are worse than you think.  
Hotwire Fision Remote: Worst Ever?

Problem 1.  The hardwired buttons are vital to using it. But they have this tiny tiny print on them. Plus, they are awkwardly placed and hard to remember. So whenever you use the remote, you have to turn all bright lights on in the room to see the tiny print.  Since the remote is often required to turn on the lights, this results in people taking out their phones to use the flashlight to see the remote. 

Problem two, for anyone over their mid 40s, using the remote with that tiny print means having your reading glasses on.

Problem three The touch screen on the remote starts out being pretty interesting. The icons are large and bright and clear. They are so bright that once your eyes get use to seeing the icons, you can't see the lightly printed physical buttons.

Using the remote means shifting your eyes constantly between three areas: 
- the big TV screen where much of the navigation happens
- the small screen on the remote which is also vital to the navigation
- the tiny buttons on the remove which is also vital to the navigation

I know that sounds stupid but it's true. Here for instance are the steps to go thru to watch a program on Amazon prime.

1.    Locate the Power On button. It's a physical button on the remote to turn something on.

2.    Now the little screen on the remote lights up. In my case, it is asking if I should "Turn off Room"?  No, I want to turn on the TV. Does it always light up to a different place depending on who was using the remote last and for what?  Does it reset after some amount of time or does it just keep whatever arbitrary state the last person who used it left it in.   So I touch the little touch screen to "Cancel". But now what? This turned off the little screen again. So now I've gone around this with the power button and this screen about 5 times when I decided to write this piece.

3.    I now see that after I hit cancel, the screen lights up with four icons: Watch, Lighting, Shades, comfort. But it only stays on for a bit and I need to act quick.   I'll click on Watch. Good, I now have a choice between Hotwire Fision TV and SmartHub.  Since they know that I'm going to the TV, why isn't this choice on the Big TV Screen instead of making me stare down at the little touchscreen? duhh? So I'll pick Smart Hub.  SHIT, I took too long. How to get the options back?

4.   I picked power again but this time, it gave me the choices that I want:  Hotwire Fision TV and SmartHub. Hurray!  After a year with this remote, I still don't know how to control  what choices the power button will give me when I use it to turn it on. I'm now guessing that it has to do with whomever used the remote most recently. So sometimes I hit the power button and all I see is some controls for the AC. Sometimes it only shows other stuff which I have no idea what they are about. I have not figured out how to get it to go from where ever it is back to the choices about TV.  On the remote, there are a Back and a Home and Menu (three buttons) as hard small little buttons on the remote but, when it's in AC and other modes, these buttons do really weird things.

5.    I hit power again, then Smart Hub. But the TV did not go on. I don't know why. Usually it does. So I hit power off again and tried again.

Again, I'm sick of this. I'll do what I usually do which is either switch rooms and see if one of the other TVs can be coaxed into showing me what I want to watch. Or I just go read.

If I get around to it, I'll write about the incredibly finicky process of trying to pick either Netflix or Hulu or Amazon Prime.  While it appears for a second to be a simple choice, the problem is that it keeps kicking into some sort of universal app and I have no idea how to use it or how to get out of there. Sometimes the Back button (the hard button on the remote) will get me out of the univeral app. Then, if I'm careful, while staring at the big screen, I can coax the indicator to get out of the default pick-a-video menu that it is in, down to the pick-an-application mode, pick it, and then go. 

Mind you, neither the Home nor the Back button are of any use in these areas, they just do weird things. Often, the screen has four big colored rectangles across the top and I've spent a lot of time trying to figure out how they work. I've recently decided that they are just distracting decorations that some idiot added thinking that while people puzzled out how to use things, a few colored useless buttons might help.  

Saturday, March 28, 2020

How Facebook Decides What to Show You

Sometimes when I'm on Facebook, I see these posts about the secret to seeing more of your friends posts.  They seem silly to me but many bright people repeat these things. So, hmmmm.

So I thought I'd share what I think I know about how Facebook decides what to show you.

My hope is that people will stop believing in what sounds to me like superstition and magic and start thinking rationally about it. Maybe, this breakthrough will light a fuse and all of the public discussion about tech and algorithms work will start getting more sensible.

Let's start with how Facebook decides what to you show you.

If you only had six friends on Facebook, Facebook would probably show you all their posts.  But, once you have sixty or six hundred or six thousand friends, Facebook must make lots of systematic decisions on what to show.  A good starting point to think about this is to ask yourself, how should Facebook decide what to show you. I'll start with three questions.

One, how does Facebook decide which of your friends you are interested in hearing about? Let's assume that  you obsessively visit one of your friend's Facebook pages, do you think Facebook should take this as a signal that you are interested in that person and show you more of their posts? Of course, the answer is yes. Similarly, what if you like, comment, and share a large number of a certain friend's  posts? Again, Facebook will take that as a signal of your interest in seeing more from that person.

What about the opposite, ie some friends whose pages you never visit and whose posts you ignore? Should Facebook take this as an indication that you're not interested in those people and show you less of their materials? And of course, the answer is that FB does note your indifference and tend not to show you their stuff.

Second question: do you think Facebook should show a preference to show you the most recent posts over ones that are a week or a month or a year old?  Are you interested in "last week's news?" Probably not.  So there's a time decay concept that FB follows and they are more likely to show you posts by your friend if you are online exactly when they post them than posts that the same friend made two days ago.

Thirdly,  if a post by someone seems to get a lot of engagement in the form of people commenting on it, liking it, and sharing it, do you think they should show that post to more people?  AgainL YES.

As far as I know, those are Facebook's big three foundation policies in their algorithm of what posts to show.

1.  Who do you appear interested in, they'll show that to you more. This is called affinity.

2.   If you are online at 3 am and one of your friends posts at that unholy hour, FB is very likely to show it to you and the other handful of people who are on FB at that time.  This has two reasons. One, immediacy is preferred by most people.  This is the principle of Time decay. However,  FB also uses the first handfuls of people who see a post to decide how interesting that post is. If none of the first viewers engage, FB might conclude that this post is a bit of a dud of a post so it'll be shown to less people. But, if lots of people in those first handfuls engage with that post by commenting and sharing, FB now has the signal that this post is hot and the showing of it will trend upwards.
3.  So the third big rule is the appeal of each post: The Weight.

So those are the basics (I think) of the Facebook decision making.  I suspect the next set of issues that FB considers are:

1.  Ads. FB sells ads, Much of what you see is paid advertising. People pay to show you ads.  Also, on pages of businesses, FB could extract more advertising dollars by only showing it to a few percent and charging money to get it shown more broadly.
2.   Pages and groups. FB has special policies around different affinities you've shown by signing up with pages (not of people but of companies an things) and groups.
3.  Look-alikes.  FB will for many decisions reason that you are behaving and have demographics like some group and will make decisions based not just on your shown preferences but by thinking that you like what other people similar to you (family, friends, look-alike groups) have been shown to like.  I think they are particularly prone to show you what your friends are seeing since that seems to create a special type of back and forth
4. Types of posts.  FB probably does not like posts with links to other sites that leads people to leave FB.  So they prefer posts with videos that are uploaded to FB, they probably don't much like posts that include links to videos on Youtube.  Probably, FB likes original graphics on FB and other internal to FB links.  Does FB prefer long or short answer posts? This is a trick question because of what we know: FB will take its cues on whether to show you graphics or long posts probably based on your personal history of engaging more or less with long posts or with images.  Again, this is what I think. Unlike the big three at the top, I haven't really researched these questions.
5.  Outrageous stuff. There is a lot of news about how FB prefers to show you views that are slightly more extreme than yours or the total opposite since both seem to produce a certain type of reaction which results in higher engagement. But this is an advanced question that we'll leave aside for now despite it probably being the reason that the world is not getting along and why WWIII is just around the corner. It's so FB could make a little more money and incidentally, build a hateful unstable world. But let's not get into that now.

Does this help you understand Facebook's decision-making?

If you are interested in learning more, this subject is called the Facebook EdgeRank algorithm.  Many businesses and social media people obscess over trying to understand and manipulate this algorithm and there is a lot written on it.

Other interesting algorithms to think about to help understand the daily experience:

- How does Google decide what to answer your query with?
- How does Youtube decide what videos to suggest?
- What about Instagram and Twitter? How is Facebook different than Twitter in showing?
- What will the stock market do tomorrow?
- And what mood will my wife be in tonight?

BTW - if this article helped you, would you leave a comment? I see that there are about 100 visitors a day of this site but I don't hear much from you.

Wednesday, January 29, 2020

SEO Issues - Jan 2019

Your Money, Your Life - This principle is being used by Google to crack down on spammy industries in which predatory practices are rampant in the consumer world. These industries have been facilitated by the web and aggregators on the web. Examples: credit counselling, pay day loans, rehab centers, etc

Google is tightening them up by adding a manual review of sites using a Quality Rating Guide in these challenging industries in which Google screens and manually evaluates sites based on EAT:
  • Expertise - does the content have meaningful additions and direction. Are there quality links out to quality sites?
  • Authority of writers and staff. Are they listed? Can they be found on LinkedIn or other sites? Do they have expertise relevant to what they are writing on? Obviously, the first step is that writing needs a listed author or authors. Maybe also a board of advisers.
  • Trustworthyness - 
IS T4L team a good enough author? I don't think so...

Tools he uses:  SEMRush, MAZPro, and Spyfu.  He likes BrightEdge but it's expensive!!!

Newer schema. BING API. - datachecker in chrome.  Other ways of checking all the cookies and trackers on your computer?

For us, Tandem does dynamic google ads, remarketing, and critero display ads.

Data Studio set up. What is it?

Here's the Google Rating Guide.

Friday, January 24, 2020

SEO History

I aspire to writing a better SEO history than I have yet read.

Here's a good source one: Search Engine Journal on Search History.

More later...

Sunday, December 29, 2019

Self Taught Engineers

I thought this article on being a self taught engineer was great. It's by Amina Adewusi who wanted to get a career going as a coder.  She learned about it through the marketing of the bootcamp people who promote the idea that people can change careers. For this she is grateful to them.

However, as a single mom, the logistics and finances of a boot camp were daunting. So while attending a conference, she was introduced to the idea that a person can learn to code on their own.

The article is excellent addressing such things as how to find mentors, when to start interviewing, how to use the "tech tests" from interviews as an opportunity, and how to hold oneself accountable.

My experience in the tech industry confirms Ms. Adewusi's view of how possible it is. As a 62 year who has been in the tech industry since the late 1980s, I have rubbed elbows with self taught engineers hundreds of times.

Sometimes I worked for them. At Silicon Graphics in the early 90s, I worked for Carol (something) who had gone to work as a secretary after high school. I don't really remember her career story (or her last name) but she was running a big engineering organization at a premier Silicon Valley company and I think her tech skills were all acquired independently by her.

More often, I have employed self taught engineers. I ran a video game development company in London in the late 90s with about a 100 people, often more, and perhaps half of the of the tech staff were self taught "hackers",  a term we used back then to distinguish between the self-taught and the formally trained.

I've run an educational software for the last 15 years and the engineers that have most helped with the actual coding were...yup... totally self taught. We also have plenty of comp sci and software engineering graduates and many with masters degrees.  And we have some bootcamp graduates. One of our clerical staff just left us to go to a bootcamp.

Many ways to get there....

Saturday, November 09, 2019

Seth Godin on Learning: Response

Usually, Seth is pithy and insightful.  Today, he sounded naive to me. Although it post was brief. Seth said

The difference between memorization and learning

In order to learn something, you must understand it. You might become so insightful and facile with the ideas that it appears you’ve memorized them, but that’s just a side effect.
Rote memorization can be done in some fields, and you can even recite what you’ve memorized to someone else who can memorize it.
For example: You can’t learn alphabetical order, you can only memorize it.
On the other hand, memorizing anything that you’ll need to build upon, improvise on or improve is foolish. You’ll need to do the work of understanding it instead.

This is, to my mind, a very naive discussion of learning. There are some very different types of learning.  There's the building of skills such riding a bike, reading a book, or telling a story.

But, to be a successful reader, a kid must master many skills and lessons.  One skill is recognizing the different sounds, learning to connect them with the letters, and to learn to decode successfully.  However, many students fail to comprehend what they learn, on major reason is the lack of mastery of the vocabulary required to understand the book.  And while it would be great if exposure to vocabulary was enough for most people to learn the vocabulary, the fact is that most people need to encounter a word fourteen times in many different contexts over a period of weeks for the word to enter long term memory.

Stay tuned, more later.

There are two very important types of fluency or memorization that should happen in elementary school.

1. Learn your time tables - I recommend Time4MathFacts as Gamified and Effective.  See this math facts educational article

2. Retain your vocabulary. While learning vocabulary has a number of clever techniques, there's a different question of how to retain. Research says that for most of us, 12-15 touches with a word, different contexts, different ways over 3-5 weeks provides a very high probability of long term retention.  Read about vocabulary retention.

Friday, September 27, 2019

Designing a Multitouch Attribution Model for Online Marketing

Multitouch attribution models are replacing first-touch and last-touch methods of attribution.
Multitouch attribution is used to figure out how the different marketing touches contributed to a customer sale. Multitouch is growing despite the complexity since it both maps to the reality of:

  • the customer shopping experience 
  •  the vendor’s activities and expenses such as PPC, SEO, email marketing, retargeting, website materials and demos, and social media interactions. 

Multitouch is complicated and there are a choice of models. The choice should be based on what you believe best maps to the customer shopping experience and what decision the company is trying to make.

At the end of the day, each company needs to figure out how to handle attribution.

At the start of the process, a key step is a strategic think on what type of analysis or decisions is being made.

For example, a company might build the attribution model to consider increasing or decreasing the investment in different methods of making contact with customers such as PPC ads on search, PPC on social, banner ads, buying lists, or retargeting. In this analysis, the company’s marketing automation system website with demos and white papers is a fixed cost that is not really up for review.

Here's a quick overview of attribution models, thanks to Jimmy Shang

1.Linear attribution is the simplest. Each touchpoint gets an equal percent of credit. So in a simple example, if the customer, a) clicked on an ad, b) clicked on a retargeted ad, and c) clicked on an email and then bought, each of these three would get1/3 credit for the win. Of course, it’s a little more complicated than this since the interactions on the website could also be included.

2. Time decay gives more credit to the touchpoints closest in time to the conversion. For example, the last email before a purchase/conversion is given more credit than the first organic search*.

3. Position-based / U-Shaped is a hybrid between first- and last-touch attribution. This method puts more weight on the first and last touchpoints, assigning 40% credit to each, and splitting the remaining 20% between the touchpoints in the middle*.

4. W-Shaped credits the first touch, the point where a visitor becomes a lead, and the final touch each at 30%. It divides the remaining 10% among any additional touchpoints. Some advanced multi-touch attribution models leverage machine learning to assign partial or incremental credit to predict the value that each touchpoint added*.

*Thanks to Jimmy Shang of Ad-Roll for his  spectacular article on attribution models. The second through fourth models are directly quoted from him.

Thursday, September 26, 2019

BAD Customer Feedback Survey Design

I just had a frustrating experience with a chat window with a vendor.
At the end, they popped up a survey and in a fit of good naturedness, I thought I would take the time to give them some useful feedback.

It was a long survey with 10 questions and a place at the bottom to write in the box. I scrolled to the bottom and filled in the box with this:

I was distracted for a few minutes and your agent disconnected me. Very frustrating.

When  I hit submit, I found that it listed all the unswered questions in red and that the survey was incomplete.  The  survey wouldn't accept the comment unless I clicked on all the  questions.

so I had a choice, fill out the rest of it or just click out.  Since I had started to give them feedback, I decided to quickly fill out the 10 questions (I picked NA for most of them, 5 if there wasn't a NA choice).

At the next submit, it turned out that my answers had opened new follow up questions which were also mandatory so I blindly clicked on them too.

What a frustrating counter productive experience that survey was.

 If they really wanted feedback, they'd accept incomplete surveys when people are trying to communicate with them.

Instead, it pissed me off and filled up their system with lots of bad random data.

Generally, I find that this is the norm. Efforts at feedback are so poorly designed, they neither give the impression that the company actually cares nor gathers meaningful data.


Friday, September 20, 2019

United Airlines Communications Policy: What?

Dear United Airlines,

We had a trip that went horribly awry yesterday due to bad weather.  My wife got separated from her suitcase.  She had a first class cross country ticket.

Let me digress for a second and say that over the last few days, I have been receiving updates from Amazon (and others) about some items that we ordered. These are items cost mostly between $10 and $50. They give us text and email updates as they ship, progress, and get delivered. It's a pretty simple IT operation which provides customers with great service so they know what's going on.

So here's what happened.  She had a flight from Ft Lauderdale to Houston (UA 2148 on 9-19). And then a connection to Albuquerque. Houston was having a huge tropical storm but United took off from Ft Lauderdale anyway. Sure, they could have rerouted her before she left but United didn't.

The plane circled Houston for an hour or so and then went to San Antonio where they kept the passengers in the plane with no food or meaningful updates for over three hours. Finally, they let the passengers deplane but gave them no instructions about what to do next.

No text updates about what to do, no email updates, no phone calls. Doesn't United have a system? They must have disruptions and weather problems of significance weekly but they seem to react to them as if they are surprised.

The United counters were mobbed, hours of waiting if they went that way. United wasn't taking phone calls, ie there were 90 minute waits.  So my wife got herself to her destination by taking a 5 hour Uber ride to Dallas and booking a flight on American to her destination.

Meanwhile, what happened to her bag?  We've called in many times and each person has been agreeable and helpful. No complaints really about the staff.  And, the United staff have been able to login and tell us where the bag is. So last night at midnight, the lady at the United baggage claim said the bag was still in San Antonio and tried to tell us to file a claim with the other airline that she finished her trip on.  Apparently, that's policy. We disagreed saying that United should get us the bag. It was a first class ticket, they provided no way for her to continue her journey, they provided no info on how she could get her back etc. In any case, we filled out paperwork on the bag and where we wanted it delivered.

We called again this morning and found out where the bag was.  We called just a few minute ago and found out where the bag is. We provided all the same information over the phone and they agreed to send the bag over.  Very nice people.

But really, should we be calling in to get information that should be sent to us?  Shouldn't the app and website have a "trace my bag" feature built in for their travelers? This sort of IT infrastructure for convenient communications seems so simple to build in this day and age. United knows their travelers, has their cell phones, and knows where the bag is and what the plan is.  Why do they believe in having people call in to get information instead of pushing it out?

What's really astounding is not just the lack of communication about the bag but what about helping the travelers? They ditched her halfway through the trip, on a first class ticket, and never followed up in any way to see if she was taken care of or needed help.  Really?