Thanks, Marshall, and I shall definitely be looking at your book!

Well, there's a math of Goldilocksness -- look up "optimization theory" and "operations research."  But it's hard and it works best when applied to cases that are well understood -- and so far, marketing is not.

bye for now.

yeah, can't keep up with the banter here - was too busy writing to really read what everyone was saying, and need to drop off now - but I invite all of you to continue the conversation with me - and hopefully, take a look at my book if you get a chance.   

I don't know if there's a proper equation to find "Goldilocksness"

Yes, you're absolutely right! Especially regarding "cause". This happens a lot in the stock market though - you can never isolate variables or do repeatable experiments - that's why its a bit of an art compared to what scientists are used to, I guess. Thanks again John for sharing this excellent point. Thank you again Marshall for sharing your helpful answers and insights! I see your point about the tachometer - but what about just bringing in more revenue than they are currently bringing in, even if you sometimes miss? It may not need to be perfect.  Exactly - "Goldilockness" - I like the term!!

Now there's an idea for another blog;  how do you measure "Goldilocksness" (i.e. not too little, not too much, but just right)?

There were cars for a decade before Henry Ford saw that they could be a useful part of ordinary people's lives instead of a fun toy for rich people.  And then it was another generation before Kettering figured out what would be needed to make them drivable by everyone, and Sloan figured out how to sell it (and they created GM).

Also the right answer may not be a maximum or minimum.  Your manual shift car always speeds up when you give it more gas (at least till it stalls) but for efficiency you want to be just ahead of the peak on the power curve, which is neither the high nor low point on the tachometer.  Similarly, some things can be marketed right into being over; you need a metric that can see that danger.

Seems incredible -- people saw the potential, but not the means of broadening to the general public?

And thanks all for jumping in with your great questions. McGraw Hill has been gracious enough to provide copies of the book. So we will be drawing among our participants for a free copy!

Noelle, the problem is that if you're trying to steer by it, you need to know whether you caused it.  If my Christmas tree business has a big leap in December, it might not be just a matter of my brilliant campaign & message in November. 

That is very sad John! Just think of the number of ideas that are perhaps still "shelved" that could be creating value!

Beth, not to digress too much, but there was a working TV broadcast setup in Philo Farnsworth's barn in 1926; the right bankers could probably have had TV out there and available before 1930.  Microwave ovens were being used experimentally in some restaurants in the early 1950s.  Everything needed to launch a satellite -- if it had been assembled right -- was available before WW2. 

Thank you!

Thanks for the informative session today

Marshall, thanks so much for participating. 

Thank you! This was very informative.

Thanks everyone for your questions, I will try to answer them offline - you can reach me at @webmetricsguru or now.seo@gmail.com and I'll try my best to get back to you, when I can.

We've hit our hour mark, and I know Marshall has to take his leave now.

Thanks - I'll check out ARF!

as equal partners to sales and biz dev, not subordinates, as they are, today, for the most part.

Thanks John. What about settling for something realy simple like up vs down? Is this not informative relative to what is already out there? Thanks.

Basicly, the biggest thing one could do today to improve results is get everyone one over to the ARF (Advertising Research Foundation) and read the Listen First! book by Stephen Rappaport - just to realize that many of the issues with social data would be easier to handle if the right people with right orientation and training were running these systems and also running the projects.

Noellel, the bigger problem is that if it's a Markov chain it's a matrix to the Nth power, not some simple coefficient.  So instead of saying Before Plus Campaign Equals After, and thus coming up with a constant to multiply Campaign by (spend X bucks, get Y more sales), you've got to understand a much more complex relation, which takes more data, more time, more insight -- and a lot more sales effort to get people who really just wanted a magic number to listen

that is essentially want you need to do to make your data set more predictable and less noisy.

Howver, by doing these steps first, most of the platforms will surfice for market research

- CUSTOM SOURCE IDENTIFICATION (for all use cases except Crisis Management)
- CUSTOM CRAWLING OF DATA SOURCES 
- CUSTOM DATA EXTRACTION
- CUSTOM ANALYTICS PLATFORM
- CUSTOM ANALYST PAIRED WITH PLATFORM with industry expertise
- CUSTOM REPORTING AROUND KPI's chosen by firm and client

But I think the platforms we're using should be more directed to Market Research, even though they platforms aren't yet good enough, in many cases to do good market research.

while marshall is answering that, wow -- john -- 20 to 50 years before they did. how can that be possible? 

Well, Noellel, about campaign effectiveness - I will say that it is important - but Social Media Analytics has the other use case, Market Research, which is actually much more interesting to work with.  Campaign effectiveness is just an extension of the same kinds of measurement your already doing in other channels - and to the extent your tools and platforms get better - that more planning is done - this will improve.

marshall --john and noellel are talking re her question: So why is campaign effectiveness measurement currently problematic? Can you run a marketing campaign and then measure sentiment, for example, using keywords from the campaign to see what people are saying onf FB etc? Is the data not clean enough? I can see why the few dollars spent is problematic for development of any capabilities! 

any thoughts to share?

Getting business to pay for development, especially long term, is a problem for every innovator, all the time. Based on casual research I've done for books I'd guess that most innovations could have hit the market 20-50 years before they did. 

Your point about the numbers is very helpful :)

Thanks John, that is very helpful. What about comparing numbers like : sentiment BEFORE campaign and sentiment AFTER campaign? Is see what you are saying aboutthe Markov chain - I think a TV ad takes 3 exposures to reach its audience which is associated with a time frame, so couldn't you just plot sentiment to check back after this time had passed to see if there is any response? Even if the shape was not linear, does this matter?

Kiyosaki pointed out that millions of people could cook a better burger than the McDonald brothers but apparently no one could come up with a better business model than Roy Kroc.  In marketing metrics, anyway, nearly all of us are good cooks wandering around looking for a Roy Kroc to organize us and sell our product

that's why I wrote the book

Ah. So the pervasiveness, easiness of access, and to use is translating into the same sense from the business side. learning from it, measuring from it, etc., should be as simple as tweeting, so to speak? that is a problem!

All these things are possible with social data, but take a lot of work because the data is largely unstructured to begin with.

Going back to establishing value of data and those who work with it: I see these issues particular to people, organizations and firms who haven't yet transitioned into networks thinking and consequently lack appreciation and expertise in strategic thinking. If this capability is present, the empowering value of data is obvious.

But, in truth, this is a factor here - we have all this work to do - up frount, before we can even use social data,for almost anything - and no one, for the most part, wants to pay for it.

BUt.. did you know that it took people several thousand man hours to figure out how to strucutre a franchise like that - any franchise - and any particualar store or outlet has a lot of work to go on behinds the scenes - customer doesn't see that, doesn't care

i'm interested to find out where you're going with this one, marshall!

But I want to point out something that few have considered - when I go into a McDonalds, and order my Big Mac, ai want to get it right away - nothing should stand in the way of me, as a consumer, getting what I'm there to purchase.

Another is that campaign effectiveness isn't one number -- it's probably something more like a Markov chain -- so there's a lot of stuff inside the black box to sort out.

I like that imagery, John -- White Coat Fetishism (plus your princess and frog from before). 

As a result - they have a problem with this "social data" that they may not have for BI, EMAIL, and House List data - they don't yet see why they should pay for it - esp as it often does not answer their basic questions - and I have to agree - it's hard to ask money when your basic platforms, even t he paid ones, can't answer basic quesations about "reach" or who is being reached, stuff you get from the others.

One big problem in campaign effectiveness measurement is that there's no large public established base of experience, so we don't know whether any given number is a big number, a little number, or a meaningful number at all.

THough Google Analytics is a powerful platform that benfitis from customization - it's also true that people feel they can do most of what they need for themselves.

So why is campaign effectiveness measurement currently problematic? Can you run a marketing campaign and then measure sentiment, for example, using keywords from the campaign to see what people are saying onf FB etc? Is the data not clean enough? I can see why the few dollars spent is problematic for development of any capabilities!

It's also true that Google Analtyics, which was Urchin (and used to cost 300 per month per account, etc) has been free since 2005.

There's also a lot of "White Coat Fetishism" -- the analyst tells them what the analyst can glean from the data, which may be very little, and they then announce that "you have to do it our way Because We Gots A Analyst. Sometimes we think we're providing guidance and they're using us for cover.

it's also true there are  alot of free tools out there, particularly around Twitter Analytics - but also for Facebook, that people are wondering why they should pay for anything when the data and some analytics platforms are free.

I spent some time in my book explaining that social data is free - we all tweet and post to facebook and blogs, and flickr, etc, and linkedin, more or less for free. 

On top of it, the clients, both internal and external, only want to spend a few dollars on it - and that is a discussion of it self.

Another is market research - few of the people who operate listening systems, esp in Marcom, but anywhere really, have a market research background - the systems can not handle large queries, they do not do well cleaning data, they can not auto catgorize well - yet, and yet the ask requires many of these things

This goes to my earlier question.  What incentive do the business people have to pay attention?  Absent that it's going to be a tough road.  Or maybe more appropriately how do you articulate the value to them so they will pay attention?

Hi Noellel - one of the most common things is the effects of their marketing campaigns - or a specific placement in a marketing campaign.

What types of things are the marketing and biz dev folks asking for that the systems are not (yet?) built to deliver?

That's why I wrote my book - for that and one other reason - to help buisiness understand people like me. 

It's also the wrong people making decision on platforms they can not meaningfully evaluate - that is what got us in is delemia in the first place.

The people making decisions are sales and business development people, who could could not care less what we do - only what they can deliver - and I find in organizations, that sales and marketing do not yet understand these platforms and tools and askiing of them things they are not built to deliver - and at the same time, can not articulate their quesations or qoals in a way that could be answered by these systems and process.

Hi Beth, it's not so much a new skill set - it's an new respect for the Analyst - the analyst, coder, whatever you want to call us - is still treated like a clerk - glorified, more or less, often well paid, but a clerk none the less.

AI is comuing to the resuce, has to - queries are getting too long and coplicarted, platforms can't run then much longer. the requirements demand AI - but most platforms simply haven't yet "gotten it" or cought up.  It's a difficult problem to handle, no doubt. NO universal anwers yet.

so a bit ago i had asked about "robotic training" and whether it requires a new skill set within an analytics team. but perhaps this need is being addressed out in the market and developed into tools?

Sounds like a very short courtship

Another example is ITA's NeedleBase - Google, I postualted, bought ITA because of it - I can't prove that, but the patent for Needle was filed in June 2010, and ItA was first aquired by Google in July 2010, the aquisition finalized in March or April, can't recall which

Experience is SO critical! Links to meaningful data :)

but till I can touch these things myself - I tend to just refer to them - when I can say to you that I have used Glide and this what it did for me - I will push it everywhere.

oh, there's my answer on glide intelligence!

Thanks Betsey!

And, though my friends at Glide Intelligence in London are on it, I can't say I have actually seen their platform run, something they've been promosing me for a while.  But from what I have been told, Vertuli Login and the AI that has been built into Glide Intelligence can handle things well such as intelligent tickers and root cause determination.

Glide Intelligence? Is that a company?

I like that question, Beth!

work in AI to make robotics and AI more intelligent is progressing - but few social platforms have inforporated it sufficently to make it useful -those who have, such as Glide Intelligence, aren't talking much.

so "robotic training," does that require a skill type that we would normally see among an organization's analytics' team, or are we looking for new skill sets, ones really germaine to social media?

The best analogy I have is weith robotic end effectors for car assemebly - they can do a lot of repetitive work -but the setup has to be very precice - every movement needs precision for the robotic program - otherwise, the end effector (hand device) won't grap and pick up the assembly unit - a loose coupling is probably not good enough.

Most of the social data, with a few exceptions, needs to be coded, tagged, categorized by humans or by robotic means - but the robots neet to be trained, and this is fairly specialized and a requirement of it's own.

there's no point in reporting if you don't see the data intergrity - but that might involve so much work in social - work we don't usually have to do in search or email or house lists (that already are cleaned somewhat) or Web Analytics - 

I don't think that proposed metrics no one used are a waste of time; I think they're sort of the frogs you have to kiss before you find the prince of the really good number.  And understanding the froggishness of the failed measurement can often help focus the effort on constructing a prince next time.

Beth, it is a big issue - cleanleness of the data is the most important issue of all - if you have no confidence in the data your collecting - your report will suck

Correct John.

Marshall, kind of sounds like this data enrichment might require a leap of faith about the "cleanliness" of the data you're pulling into your data warehouse. Or is that a nonissue?

IN another example, you can take web analytics data, if it is set up well, as I discussed yesterday, and add in the social data that could have referred back that specific page, and then tie in you social efforts to tarffic on the site.  Also, you could have used certain types of tools that checked browser hisatory to see if you have been in certain locations onteh web web, such as microsite - so that your branding (if you ran a campain) is effective or not, abnd by how much

It sounds like a lot of the current work in relational databasing and table-driven AI will probably eventually contribute a great deal to the triangulation process.

wonderful!

In turn, once you have that data, you can use zipcode information (that you might have) to trigaulate geo-demographics and overlay that information to your list - this give you a much richer data warehouse to make predictions on future behavior of these people.

John, as a result of measures going viral but than that fade out ... is that mostly  wasted effort, then, or can we still get value from those "experimental" measurements?

Hi Beth, triagulation is overlaying data that adds value - for example if you have a database list internally, you can do more with it if you use data enrichment (aka PeekYou.com to get twitter and facebook account information) and so on - as well as sites that you never collected in your database

Thanks for that information regarding the Web Science Trust. I will look into this as well.

Some more established branches of statistics also have the tradition of the journal article with a title something like "a proposed measure of ...." or "a way to measure ..." A few measures go viral and become standards (like P/E ratios, GDP, average hours worked, etc.)  Most fade out.  Eventually people figure it out.

so let's move onto next question: Marshall, we've gotten a couple of questions relative to your comment about triangulation. can you explain more about triangulating analytics platforms in regards to social media analytics?

i'm not familiar with the web sceince trust. will definitely ahve to look into that! 

The Web Science Trust, I feel, is the standards body for the future - it is not about about marketing or PR, like the others - but pure science.

FInally, the Web Science Trust in the UK, and here at NortthWesten University - and some others, I feel, is the most exciting, "sexy" body to be part of - it's headed by none other thanTim Bernes-Lee and his cohorts and operates as a "trust" to study web science.

Also, the Web Analytics Asssocation is doing some measure of work in this area, but they did more when I was involved - after I left, and partly over this, they fell back to foucsing on marketing measurement and did not do as much as they could have.

Hi Cordell, how do we faciliate creating standards? We work and empower the IAB, the CIPR, etc, standards bodies to do that for us by parcipating in them - they already have study groups that have been going on for a while and are doing worthy work in this area.

so the idea of creating standards for this is interesting, as cordell asks about. anything more on that marshall?

Thank you for answering my earlier question Marshall :). If you can elaborate - what is the best/easiest way to triangulate? I believe that Aglevy asked a similar question also.

But I understand many platforms resist standards because they inhibit innovation and are somewhat forcing transparancy, which may reveal more than what the platforms would like others to see and know abourt them.

I think you just persuaded me to buy that book!

And the problem with a good universal framework is that we'd need someplace doing a long term commitment to basic research to develop it; it's too new for the universities right now, and the innovate private labs like Bell, Westinghouse, IBM, etc. are sadly fading from the picture as the immediate-bottom-line considerations take over.

John, the other thing that I discussed in my book in chapter 1 was the IAB's VAST standards for Video Advertising did wonders to help monetize that sector - before VAST, there was no way an advertiser could get the right controls on a publishers site for video ads.

good question cordell (and welcome)!

I'm intrigued by the idea of creating standards in a rapidly evolving environment.  How do you/we go about facilitating that?

Hi John, yes, the basic thing that's missing is a universal framework we can all agree on, and work off of.  Think of the Network theroy that most communications is based on -7 layers of information all the way from the device level and back (ie: TCP/IP, Application layer), etc.

We need that kind of frameworks so we can create metrics and subcategorizations around it

Actually I thought the list of CUSTOM was a pretty good start on an answer; if everything CUSTOM were available in a STANDARD AND AUTOMATED version, you'd pretty much have the problem knocked.

Marshall -- how 'bout JOhn's first question: what do you find is the biggest Thing That's Not There Yet in the way of developing good, standardized, replicable social media stats? 

Hi Noellel - yes, triagulating data does help - getting your basic mentions data and overlaying to Google Analytics can be done, but it takes planning - but the more you can triagulate, where there is a good "fit" such as I mentioned at the end of yesterday's webinar with Awareness, the better the information will be. more useful, that is.

Marshall, a few questions to get to ... how 'bout starting with Noellel's?

Hi, Beth and Marshall.  I'm just thinking about the number of bottlenecks that projects like this run into.

A lot of people who work in the marketing communications space don't "get it" yet. Most don't want to pay for this, or are able to peform it, either - best situation is to partner intelligently - in this senario. 

Thank you Marshall for making yourself available to answer questions via this chat. I have a quick question regarding social media measurement. The question is : how easy is it to convert from measuring responses or sentiments by social media chanel - FB, Twitter, etc to more conventional market segment categories such as income, age etc. so that it can be compared to an integrated with product purchase predictions based on these categories, for example? Are these a part of teh avilable custom reports you mention?

Yesterday in a web conference you discussed triangulating analytics platforms. Can you discuss further here?

Hi John

Basic things to do it right include

We have to separate use cases first between PR/Comm (cover that in the book, Chapter 7, I believe) to Market Research to Enterprise Business Intelligence, etc - there are a few basic use cases and often, hybrids that are more problematic to set up and maintain because multiple stakeholders have different expectations and often want tailored reporting, often requiring separate reporting (or redoing reporting over and over again for different stakeholders).

Hi, glad to join the crowd ... what do you find is the biggest Thing That's Not There Yet in the way of developing good, standardized, replicable social media stats? 

Excellent quesation again, Beth, here's what i found works the best in this space, given that we're talking about 2011-2012 - if we go past that -things might change a bit - but this is how it is today

So, you have been studying social media metrics for a "long" time.   Obviously a lot of companies are going crazy trying to figure out how to best mine data from social sites for business intelligence purposes. 

And finally, if people want to know more - there's the technorait article http://technorati.com/blogging/article/social-media-analytics-marshall-sponder-provides/

As a result, I’ve created my own view
based on personal experiences that I provide to the industry as a
whole and to my readers in particular in the book.

And, I’m also close friends with many of the leaders in this

field, many of whom have contributed to this book case studies
and various insights you will encounter in the following pages.
I’m fortunate to be in a position to see the social media metrics
industry in a holistic way. 

So you've got the pedigree, absolutely!

Cool.

And, furthermore, if I have not been involved in something

personally, I almost always know people who have been, and
I’ve wrung the salt out of their wisdom at conferences, bars, and
tweetups. 

Now, many people have read my WebMetricsGuru.com blog

over the past six years and can vouch for the fact that I know
social media, search, and Web analytics quite well. It’s gratifying
that I’m considered by the industry to be a practitioner, a
hands-on analyst, and an expert in the field. I am not a journalist
or conference promoter; what I write about is the very
work I’m involved in and have pondered over deeply for several
years. 

HI Aglevy

Hello

Also, Richard Pentin’s “New Framework for

Measuring Social Media Activity”10 defines the four A’s of
social media measurement: awareness, appreciation, action,
and advocacy. Pentin’s paper proposes defining and measuring
core key performance indicators by social media platform,
marrying soft metrics with hard financials.

Hey Beth, excellent question:

Based on my experience as a board director at the Web
Analytics Association from 2007 to 2009, where I built the social media committe, I know firsthand
that standards are necessary for the social media analytics field
to mature to the point where it can interoperate with other
business research disciplines. To that end, the UK-based IAB. 
Social Media Council has launched an initiative with a focus
on standards drafting. 

I'd like to start with a question by way of background: What motivated you to write Social Media Analytics, and what gives you the authority to do so?

Hi Folks, I'm here and let's start now

Hi Betsey, great to meet you here

And AnalyzeThis!

Welcome!

And Chisquare!

Hi Betsey! Hi Noellel!

Hello!

Hi, Betsey Merkel here. I'm a community developer with advanced skills as a Process Designer, Network Strategist and Social Media Knowledge Curator.

Hello, all!

Welcome Marshall!

Hi Guys - here now

30 minutes until start time

Testing in Chrome - all set to go, again! Here's where I'll run my chat from. See you guys at 3PM.

Testing my connection for 3PM Chat - all set to go!

Here's where we'll chat with Marshall Sponder, renowned Web analytics guru and author of the recently published Social Media Analytics: Effective Tools for Building, Interpreting, and Using Metrics