Comments
Analytics Best Practices: Plan for the Human Factor
You must login to participate in this chat. Please login.

Thanks Chris and All Analytics team.
Data Doctor
DigitArchitect, yes it is good to design processes that facilitate in analytics data generation. That's a result of good strategy
Data Doctor
Joe, there must be leadership supervision on these projects that cost significant sums.
Data Doctor

And, thanks to everyone in the audience for joining today. Have a great afternoon.

Editor

Sorry . . . deriving.

Prospector

Looking forward to it @CMazz Thank you!

Blogger

Thanks so much, Chris!  Really appreciate your insights today.

Blogger

@Chris Would you agree that there is a need to build the infrastructure of analytics at the same time as deviring the analytics strategy that somewhat drives analytics at the same time? I guess what I mean is, do you start with one business problem and then begin to build the infrastructure of your analytics data, tools, team etc. . . or should this be both a grass roots approach (i.e. working to clean data, seeking out tools for visualization or analysis) as well as a top down business strategy effort?

Prospector

Chris. Thanks again. Great presentation.

Editor

I think we've all seen analytics efforts that ended with no usable data or no action taken. Big waste.

Editor

My apologies everyone but I need to head to a meeting. Really enjoyed the conversation.  Happy to continue the dialogue and will share the info that was referenced.

Prospector
Jamescon, true. Functional managers should get involved. I just recently faced this and I got a bad flock. Then I went on to LinkedIn and posted vacancy to the right groups and I got the right man for the job
Data Doctor

In the example of the late payments they should have designed the 'intervention' mechanism first and then worked backwards.  Or they should have incentivized the customer service reps on reducing accounts receivable.

Prospector

Cannot tell you how many times I have seen tens of millions of dollars spent with little to no value.  We need to get behind the promise of analytics to driving real value.

Prospector

That's a good point.  If the analytics are costing more than the ROI, then, well, time to go back to the drawing board.

Of course, such a situation suggests poor analytics deployment/a problem with the human element discussed earlier.

Blogger

In that example of the late payments. What should the company have done differently fro mthe start?

Editor

And ask that question of the organization...how much investment and time are we spending on analytics 'production' vs. analytics 'consumption'.  Often out of whack.

Prospector

If anyone has a question from earlier that Chris may not have seen, be sure to bring it back up.

Editor

I like that 80% guideline

Editor

I like the idea of managers being more involved in the screening, recruiting process. Not only does a lot get lost in translation with the HR team, but that manager can really help to refine the search once resumes start coming in. I've worked with managers who left the whole thing (prior to actual in-person interview) to HR and the result tends to be a bad hire or bad flock of candidates.

Editor

Creating value and 'proving' to the business what is possible creates more equity and goodwill to tackle other areas.

Prospector

I see a question about exploratory analytics.  I absolutely think there is a role for this.  There are cases where truly 'exploratory' and 'unsupervised' analytics is a appropriate.  My hypothesis is that MANY organzations should have 80+% of the effort (initially) on well defined business issues / opportunities.

Prospector

Chris welcome back. In case you missed it: @chris. One thing we were discussing on the chat board is the aspect of knowing the business problem before you launch an analytics project. However, one poster noted that sometimes it's the analytics project that identifies the problem. Do you see a role for an "exploratory" analytics project on occasion, versus the more typical "operational" (problem driven) project?

Editor

@Chris - Look for James' most recent question.

Prospector

The history sometimes takes a while to load.

Blogger

Thanks Chris... Good talk

Sometimes, in some specific fields, you need to have a type of SME to help provide specific subject matter information for the "team" to work on.

Prospector

I seem to have gotten knocked out of the chat.  Back now but cannot see the history.  Sorry if I missed some questions.

Prospector

@chris. One thing we were discussing on the chat board is the aspect of knowing the business problem before you launch an analytics project. However, one poster noted that sometimes it's the analytics project that identifies the problem. Do you see a role for an "exploratory" analytics project on occasion, versus the more typical "operational" (problem driven) project?

Editor

Indeed, it probably speaks to organizational and departmental culture quite a bit -- analytics or not -- when the managers are very involved in recruiting versus little to not at all.

Blogger

Also, there is a kind leaning toward the 'unicorn' as building a team often costs more than just hiring the "one ring to rule them all."

Prospector

I doubt that they do.  My experience is analytics 'leaders' need to be VERY involved in recuriting.  Often a big part of the bob when teams are in 'build' mode.

Prospector

Chris, I think we all agree that crafting a job description around some of these things is tricky. But I wonder if the hiring managers "know the talent/skill when they see it"

Editor

We will have it posted on ey.com as well as Forbes (our partner in the survey).  Will get it out to this audience as soon as we release it.

Prospector

@DigiArchitect we haven't any emoticons here, though you could suggest it to the powers that be.

Blogger

Thanks to Curt, Chris, and Jim for this! :)

Blogger

Would be interested in the group's perspective on my fundamental premise...that many organizations are not thinking enough about the 'human element'.

Prospector

Thanks @Chris. Do you have a URL for the report on the survey you said we'd be able to see?

Blogger

Thanks, Curt and Chris!

Prospector

I see many questions about job descriptions...this is tough to answer generally as it really gets to being very clear about the various roles that are needed.  I have seen lots of examples of job descriptions that are looking for 'unicorns' as opposed to thinking about building blocks to a team.

Prospector

I wish there was a laugh button. Love the comments here!

Prospector

Nice job Curt and Chris. Great stuff.

Editor

@Jamescon: A certification that hasn't been invented yet, probably.

Or the certification offered by the HR Director's brother-in-law's company.

Blogger

Ha, I can type, just can't talk.

Editor

Good one, Joe. I've dealt with that same HR person!

Don't forget, what certification does that person need?

Editor

@Joe I really love seeing jobs called entry level that want 3 years expereince. One refers to the pay scale and the otehr to the skills they want.

Blogger

How do you get a champion at a high enough level to promote, support, and make progress in developing out the positions and garner resources and cooperation to include analytics in a business strategy?

Prospector

HR: How many years experience should the applicant have?
MANAGER: Ah, well it's not so much about years of experience as it is about the skill sets and the person fitting into the organization's culture.
HR: Right, but how many years, if you had to quantify it?
MANAGER: Well, they should have *some* experience, probably.  But they don't need to be a subject matter expert...
HR: So how many years?
MANAGER: A few, I guess.
HR: So, like, 3?
MANAGER: Ah, two to three, sure.

JOB POSTING: MUST HAVE TWO TO THREE YEARS OF EXPERIENCE

Blogger

Those key words are what brings in more resumes, which HR thinks is great.

Editor

@Joe I'm sure that it does happen in some form or other more often than companies like to admit. It explains why some seriously unqualified people get their jobs.

Blogger

@James True, they like to match lists of keys words and to eliminate those who don't hit all of them.

Blogger

Answer: Hire your co-worker's cousin's friend.

Blogger

HR loves easily defined job descriptions that fit with all of the other descriptions already out there

Editor

@joe,  100% correct with the word screener not a good use for companies

Prospector

Ditto on Joe's question.

Prospector

Is there a Cliff Notes version?

Blogger

Question: How to communicate these needs -- for instance, exceptional visualization skills -- to the HR department and help make sure that these very spatially minded people aren't being weeded out by the word-screener?

Blogger

Absolutely agree about visualization talent!

Prospector

@Ariella: XKCD #1028 is very on point here.

Blogger

When I spoke to people about how they're preparing people at business schools, they said that they need that mix to fill the jobs that remain open -- a combination of bus and data understanding.

Blogger

@digitalarchitect. No doubt, we can't limit analytics to those problems that we as humans know. My concern is with the analytics project that runs off on its own in a back room somewhere without understanding what people are seeing in the customer-facing or operations roles.

Editor

Definitely a translator needed.

Prospector

I recently spoke to some people about just this. They have to learn to speak a common language, so that the business people can communicate with the data people. 

Blogger

Definitely agree that IT is necessary, but probably not the place for the primary analytics talent.

Prospector

Storytelling is ultimately your bridge between the raw, unemotional numbers and the fuzzy and intuitive "human" side of business.

Blogger

John Barnes, as I recall, has some great pieces in the A2 archives about interpreting the data and analysis to tell a human story.  Here's one on correlation vs. causation: http://www.allanalytics.com/author.asp?section_id=1413&doc_id=247352

Blogger

Agree with @Joseph. To some extent, limiting your starting point to starting with business use cases can limit the usefulness of your data and the story it can tell. You could be assuming that only those business problems we are aware of or have already identified are the only justification for developing analytics from the data you have. I'm convinced that there are plenty of hidden issues or insights that could be illiminated if you are starting from the data looking for indicators that ultimately generate new and maybe even more appropriate business use cases that need to be addressed. It seems like having a business case ahead of time could narrow your scope of insight.

Prospector

My CMO just yesterday came by my office to hand me an HBR article titled "The best data scientists know how to tell stories".  Anyone here a children's book author? 

Prospector

@Joseph. Great point about pre-conceptions. Maybe that's a future radio show: How analysts and the business managers they support can step back from their prior notions and biases.

Editor

@James: Yep.  The "Aha!  Look here!" moment.

Blogger

Good point, josephdavis57

Blogger

Yes.  Tell a story.  Define your business problem.  Enumerate the business environment factors at play and the raw data.  Use case studies related to similar situations and explain how they're the same or different from your own situation.  Then apply the data and analysis to your own present business problem.

Blogger

@Joe. The way I look at it personally is that exploratory analytics spawn a closer look and maybe another round of analytics that ends with "Here's how you can fix the problem that we discovered in the exploratory stage."

Editor

As long as you just aren't pulling data that supports the conclusion or story you already have in your mind.

Prospector

Yes, because people respond to stories more than to numbers alone. Someone even pointed to that as the reason people are not as concenred about hacks as they should be -- the threat stays abstract as numbers.

Blogger

Of course, ultimately, all analytics should be actionable in some way -- exploratory or not.  Some data and analysis are simply more directly actionable.

Blogger

Wow, that's exciting, we'll be on it first!

Blogger

Welcome, Joe (other Joe).  :)

Blogger

Good point DigitArchitect, two types of analytics: exploratory and actionable. In these cases we're really talking about actionable.

Editor

It's like Stephen Covey's admonition to "start with the end in mind."

Blogger

Only 15%? We really do have a long way to go then.

Blogger

Hello my name is Joe and I'm an actuary for an commercial insurance company

 

Prospector

It seems this is a chicken and the egg problem. There are questions that already exist that are business problems that are needed to be solved, but there are also "problems" or insights that can be discovered by actually just looking at the data and trends in the data that you have. It seems to me that there are even business problems that were unknown that can be illuminated by being able to evaluate the data using tools that help to tease out trends, correlations, and patterns that would not been seen without analytics.

Prospector

Is a report on that survey out already?

Blogger

Ariella, by contextualizing, do you mean in the collections example? I think the issue there was that you probably had people who had plenty to do as it was, and nobody thought to say, "You can help us collect these debts, and you get a share of it."

Editor

Curt's point is so dead-on here.  Technology is not the be-all and end-all.  It's an important element, but it's all about the deployment, the culture, and other human aspects ultimately.  The tech is just a tool.

Blogger

@Ariella: Indeed, the data are only as good as the people collecting and analyzing the data.

Blogger

Definitley you have to start with a defined business problem. Otherwise you just collect random data that may not relate to it at all.

Blogger

But didn't they think that a gentle reminder would be enough to get what they wanted and that contextualizing it in that way would win loyalty rather than annoyance?

Blogger

Playing a bit of catch up here with technical difficulties.  Glad that these radio shows are archived!

Blogger

Shawn, do you find that real end users get factored into analytics programs? For example, how data will change someone's job?

Editor

Welcome, Shawn. Glad to have you with us today.

Editor

OK, got audio but couldn't post. Now I got both to work.

Blogger

Anyone else having difficulty with audio?

 

Editor

Hello, My name is Shawn Denny. I am the Information Services Librarian at Taylor University.

Prospector

In monitoring analytic pros' comments on sites and in social media it seems that there is a subset (not a majority) who forget that somewhere down the line there is a human who has to put that data into action in day to day work.

Editor

I head the recorded intro bu tnothing after. Here we go with more refreshes.

Blogger

For folks who are new to A2 Radio, if you don't hear the broadcast when it starts at 2 pm (ET) try refreshing your browser. That usually works.

Editor

It should be very enlightening!

Blogger

Chris Mazzei, Ernst & Young Chief Analytics Officer logging in.  I am looking forward to speaking with everyone.

Prospector

Just a half hour til air time. Should be a good show

Editor

We'd love to have your voice in the discussion here. To take part, just type your comment into the "Your Post" box and then click on the "Post" button below the box. Feel free to introduce yourself before the show starts -- I think you'll find that we're a very friendly community here! 

Prospector

Hey, everyone, we're glad you could join us! When the show is scheduled to start, an audio player should appear above the "Your Post" window. If it doesn't appear, you might need to refresh your browser until it does. If it appears but doesn't start playing, then you may need to click on the "play" button on the far left of the player. 

Prospector


INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +