Sunday, July 1, 2018

B2B Marketing Data

AnalyticsIQ is about to launch our foray into the B2B marketing data biz.  It is a very natural extension of our current business - consumer data -  in that humans actually make all purchasing decisions at companies.

In jumping into this business we did a lot of research regarding the 'state of the art' for this type of data.  In collecting survey data to be used to create our bespoke data, we found some interesting things:
  • Most users believe that data compilers (magically) know things like 'number of employees' and 'annual revenue' about 19 million US-based companies - they don't.  In fact, we found that these attributes on the leading compiled files had only a weak correlation to a sample truth file.
  • There has not been a lot of innovation in this space.  We find that the standard compiled data is highly commoditized and that other data consists (primarily) of scraped data with limited reach and reliability.
As we have done with our consumer data file, we have attempted a fresh approach.  To get maximum value from our consumer file we have invested heavily to develop as many 'links' between consumers and their place of employment as possible.  We are currently at 55 million confirmed links (other providers, we have found, throw out big numbers for total consumer links - but you have to consider the possibility of really loose matching to get to that point.  Just saying.).  

We are also building accurate and validated tools that predict the following:
  • Level of purchase influence for each company contact (decision maker, influencer, none)
  • Preferred content for each company contact (white papers, webinars, etc.)
  • Preferred contact method (email, phone, in-person, etc.)
  • Primary purchase driver (relationships, price, quality, etc.)
  • And much more...
Look for our announcement regarding our launch of BusinessCore later this month.

Friday, April 13, 2018

Ethics in the Data Business

I know - the title of this blog perhaps made you laugh a little.  Kinda like saying 'technology in the outhouse business' or 'environmentalism in the oil business'.

It is true that the marketing data business draws particularly unsavory characters.  Having jumped into this field after a career in risk analytics - I have been shocked at the pervasive lack of concern for the real humans represented in our datasets.

I am particularly disgusted by the profit-based obsession of the marketing data business.  Don't get me wrong, the business purpose for any company is to make profits - in part.  But equally important is an acknowledgment of other stakeholders in any business:  employees, customers, the environment and society at large.

Here is what I personally believe that should be required for any data company CEO to sleep at night:
  • A realization that we have a responsibility to police how our data is captured.  'Grabbing' data off of the internet does not meet this standard.
  • Empathy for consumers impacted by our data.  Working with predatory marketers (while perhaps legal) is unacceptable.  If you would not let your elderly parents/grandparents buy a particular product, why are you helping to sell this to others?
  • A very clear mechanism for consumers to 'opt out' of any marketing database.
While people who know me well would be shocked by a biblical quote, here it is:  Do unto others as you would have them do unto you.  A great philosophy for running a data business.

Thursday, March 29, 2018

Third Party Data on Facebook

The backlash against misuse of private data has started.  Yesterday, Facebook announced that (effectively) no third party data will be used to target advertising on Facebook.

While there may be ulterior motives for this move - it is clearly an overreaction to a real problem.  As I have previously mentioned, it is an unsustainable violation of consumer privacy to gather internet behavior to target advertising.  Of course, our friends at Cambridge Analytica are the new poster children for this practice.  What is surprising is that they are banning predicted data from their site.  As a refresher, predicted data makes a variety of predictions based solely on public data.  So, AIQ's 'Facebook Influencer' model actually contains no Facebook data - we leverage analytics to make accurate predictions.

Although I have always believed that the future of targeting involves heavy analytics, my faith did waver a bit over the last several years.  It did seem that 'internet scrapers' were having a lot of success and no one was especially concerned about the privacy implications.  The world has turned.

Wednesday, March 21, 2018

Cambridge Analytica

I am very proud that my firm, AnalyticsIQ, was named one of the '10 most innovative companies' by the Technology Association of Georgia.  But...it is frustrating that it was necessary to spend a portion of our acceptance speech contrasting our approach to the other 'analytic' firm in the news.

The reality is that the need for analytics would be greatly minimized if we just grabbed/scraped individual Facebook profiles - I would argue that we know what people are all about at that point, without the need for prediction.  AIQ does have data that is analogous to some of the CA data - such as political preference, etc.  But here is a major difference - we have used hard-core analytics to make these predictions about individuals based solely on public data - and we independently validate every one of our predictions.

Of course, 'knowing' that someone supports a particular candidate (or cause) is undoubtedly better, but not at any cost.  There is a major backlash coming - led by consumers against Facebook and other social networks.  I am more than OK offering 80% solutions that are 100% on the right side of where the world is headed.

Now dismounting my high horse to have a glass of wine....

Wednesday, March 7, 2018

Deep Thoughts from RampUp 2018

I am just returning to the ATL from RampUp 2018 in San Francisco.  It was an intense couple of days.

First thought:  Despite my concerted effort to look and act like i am 23, I actually felt very old.  I was born 53 years ago, which makes my age 136 in 'tech years'.    That aside, it is great to interact with so many super-smart kids, who seemed genuinely surprised that i could remember what i had for breakfast.

Thought 2:  I am shocked by how many startups are focused on geo-location data.  Don't get me wrong, it is very cool to know that someone spent their afternoon at a bar before going to church, but i am truly surprised that 20 companies have gotten funding to pursue this angle.

Thought 3:  I am bullish on Liveramp.  They really have a cool business model and have been a great partner to AIQ.  While i still maintain a few concerns about the near monopoly they have in the on boarding space, they at least seem like benevolent overlords - which is all we can ask for.

Thought 4:  Digital marketing is now officially just marketing.  I know that some of us still sell buggy whips, but all money, energy and intelligence is focused on digital.  The new world is Tesla, not Pontiac.

While I still expect Skynet to activate before RampUp 2019, it is good to know that humans still have a small place in the marketing biz.

Thursday, February 22, 2018

AIQ is hiring! Senior Analytical Consultant

AnalyticsIQ is seeking a highly qualified individual for the following position:  Senior Analytical Consultant.

This is highly visible position reporting directly to AIQ's COO.  The role requires providing technical support to AIQ's sales team and technical management of existing accounts.

This job requires 5+ years of hands-on analytics experience and 5+ years of sales support.

Salary and title commensurate with experience.

Please respond with resume to gailm@analytics-iq.com

Wednesday, February 21, 2018

The Case for Automated Modeling

I have been a vocal opponent to all forms of (truly) automated modeling for years.  When I started my first analytics firm (Sigma Analytics) in 1998, there were some seemingly valid players out there.    Before we went out and hired what turned out to be 20 humans to build models, we did take a look at these automated modeling platforms.  None could touch what a well trained human could do.

Just to be clear, I am not innately against automated modeling - it is just that we have tested most newcomers to this field and have universally found that an experienced modeler can outperform automated platforms.  By outperform, I mean in predicting the future (as opposed to what automated platforms have traditionally done well - predict the past).

My view has recently changed - we have tested a platform that equals the power of an experienced analyst, and does it far more efficiently.  Of course, the ideal solution is for an analyst to intervene into a 'near final' solution to add in the art of modeling.

What does this mean going forward?  I think it will allow experienced analysts to cover far more ground (maybe 3X), while still producing really solid algorithms that actually predict the future.  I am not under any illusions that this platform will remain a competitive advantage for my firm - but it does mean that we need to continue to adapt new technologies as they come along.