I'm a Senior Systems Engineer at JMP.
Today I'm going to be talking to you
about quantifying your organization's analytical maturity
in order to make data- driven decision making
so you can do things better.
I've been using JMP for some time.
I actually learned JMP on a boot leg copy
of version six when I was in graduate school.
It's been a part of my life for a long time,
and being able to use it to help others find success is one of my great joys.
Today, I have a lot of people that are helping support this,
primarily is Brady Brady.
He is a Principal Systems Engineer also at JMP.
He helped craft the background tool
to perform all the analytics that I'll be presenting
and that you may be able to take advantage of.
Then, of course, I want to thank my team.
We work with high tech companies here in the United States,
and that is Ben Ross, who is a Strategic Account Manager,
and Kyle Bickford, a Senior Account Executive.
The goal for today is to demonstrate
how you can collaborate with your JMP support team
to quantify your organization's analytical maturity.
By quantifying that,
you'll be able to understand where people are spending their time
and how competent they are,
and then use that as a benchmark to track progress.
By working with your JMP support team in this effort,
they can help craft the support necessary
to bring your team from where they are to where you would like them to be.
How do we do this?
We use the JMP Analytical Workflow Survey.
I'll refer to it as JAWS throughout this talk.
I feel that's a little catchier.
The tool, it's a quantitative tool
to just measure the analytical maturity for your entire organization.
Despite JMP being in the name, it is not just for JMP users
and so the idea is that you assess an entire organization.
You can break it down by department, by job title, etc.,
and then use this by annual rerunning of the survey to understand
if you're moving in the right direction and what areas you need help in.
It also helps you identify white space where maybe analytics is heavily used,
so that you can bring that up to your management
or whoever it might be
in order to help get your entire organization moving in mass
in the direction that's going to drive your company
towards discovery, efficiency and growth.
How does it work?
It's simple. It's a five minute anonymous survey.
It measures the amount of time
a person spends performing a particular analytical task.
It's not just one, but all of their analytical tasks.
It understands their self-professed competency.
By doing this anonymously, we find that people tend to be more honest.
Not only do they get to say,
I spend two hours a week doing data visualization,
but they can also say,
I don't really know what I'm doing, or I got this.
I'm definitely an advanced user.
So we can understand how much time they're spending
and what level of competency they have.
Maybe most importantly,
it'll give the person opportunity to say, I need to know how to do this,
and I need help because I don't know what I'm doing,
or I need to do this to complete my job,
and I'm totally competent in this.
It's time spent,
how well they think that they're able to do it,
and where do they need support to do their job better.
If you think about it,
those three things for management to know that,
I mean, that's incredibly valuable.
This is a very easy way to identify this information,
and we present it in a way that has great visuals,
easy to comprehend and digest,
easy to share with upper management to help build that roadmap
for getting your organization from where they are
to where you want them to be.
What exactly are we looking at?
Well, if you're unfamiliar,
this is called the JMP analytical workflow.
It works left to right.
On the left you have where is data coming from?
For example, in the survey, it's going to ask how much of your week
is spent interacting with files or documents or databases or web APIs.
Your data comes into some analytical tool,
preferably JMP, but it might be something else.
Then people spend time doing tasks.
From accessing data
to performing basic data analysis and modeling.
Maybe they're doing reliability, consumer research.
Maybe their job is more focused
on building automations for the organization.
They're doing something or a series of some things
that take some time and requires competency.
Then, of course, they need to share that.
It'd be a shame if all of our hard work just lived in our hard drive,
we presented it in a PowerPoint in a meeting, and then it just goes away.
We want data to come in,
we want something to be done to that data,
and then we want data to be shared
with the entire organization so that people can learn.
The longer I am in this position,
the longer I recognize there's no such thing as a one-off problem.
Everything comes back in some shade of gray relative to where it started.
If we have an area or a way to query
all of the problems an organization has solved,
it's going to save a lot of time in the future
because people aren't going to be starting at ground zero.
W e want to measure all of these things.
Where are they spending their time?
How capable do they feel they're at doing that?
And where do they need support to do that better?
The first step is collecting the data.
This is, as I mentioned, about a five minute survey.
The first three questions
are c ompletely customizable to fit your organization.
This is anonymous, but we want to know some things.
We want to know where are these people located?
For some of you, it might be we're all in one place.
Maybe they work in the office, they work from home.
Who knows?
Maybe they work in the United States.
Maybe they work in Europe. Maybe they work in Asia.
Maybe they're just spread out in different sites across the US.
But we can customize that to fit your o rganizational design.
Then we want to know what department they're in
and what are their job roles.
This allows us to slice and dice that data once that survey data is collected
to better understand where are things working well and where do we need support?
You may have an R&D department that's in one location that's just crushing it.
They are very competent. They're very capable.
They're not spending too much time because they built automation.
Then their peers at maybe a newer location are way behind.
They're the ones that need support by designing it in such a way
that we can slice and dice it by department, by job title, by region.
We can really get to the heart of where support is needed
or really pat ourselves in the back because we're doing things well.
We are where we thought we would be.
But I've administered the survey many times,
and every single time I hear, I thought we were better,
so this is a great way to measure that.
From there, we're going to look at how much time
do people spend doing particular things.
You'll see, none of this is JMP specific.
It's really designed for anyone.
Anyone working with data needs to get that data.
Anyone working with data needs to clean that data,
put it into a position where they can analyze it,
look for outliers, visualize it, whatever the case may be.
Additionally, we'll have something very similar to this
to collect data on what their competency is.
Then, of course, as I mentioned,
we have the opportunity for them to say, I need to access data.
It is critical to my task, and I am not good at it.
I'm inefficient, I don't know how to query our database,
I don't know how to bring in 55 CSV files efficiently,
I need advanced training.
Or I'm really good at this, or I don't even need this,
someone just emails me a file and I do my work.
It allows them to tell you what do I need to be better at,
so that I can be better at my job.
At the end of the day,
most of our people want to do their job well.
They want to be successful. They want to advance in the company.
They want to show that they have value and worth.
This is an opportunity for them
to tell you where they think they need help.
There's been a few instances where I've talked to a management team,
and they're like, this department doesn't know how to do that.
Then the survey results come back and the manager is like, my goodness,
they all feel like they need to know how to do this better.
I have no idea.
This is really an eye- opening opportunity
for a lot of people when they see these results
to really fully understand exactly where their people are
versus where they think they should be.
We collect this data,
your JMP support team will analyze this data,
and then they'll be able to present it to you.
Now we're going to walk through what this data looks like,
so you can get a sense of what will I learn.
We'll go at a couple of different views.
At the 10,000 foot view,
we get these heat maps that show where are people spending their time.
I've broken this up by organizational wide on the upper left,
by job title on the upper right, and by department at the bottom.
If we just look at the upper left,
we can see that people are predominantly interacting with files,
probably Excel files or databases,
and they're doing a lot of data exploration,
a lot of basic data analysis modeling,
very little reliability analysis,
a little bit of quality process engineering,
and they're primarily sharing images.
If we look at job title,
it's a similar story, but this particular job title
is maybe doing a little bit more time running design experiments.
Then by department, maybe this is an analytics department
or a chemistry department, but they're doing a lot of DOE,
a lot of basic data analysis,
a lot of data base, and then sharing images.
As a leader in your organization,
you might ask yourself, are images the best way to share this data?
If not, this shines a light on the fact
that your company is spending a lot of time sharing images.
Maybe A, this could be automated,
or B, maybe we want to push people in a different direction
sharing some other data format, writing particular reports or etc.
But it just shines a light on the things that are going on.
Then you and the support team will work together to better understand,
they'll probably have lots of questions,
is this what you want?
Is this what you expected?
Should you have people doing more quality if you're, say, a manufacturing firm?
Do you want people to be quality minded or do you have a quality department?
Those sorts of questions will start to flesh themselves out
and they'll help you craft
how they might be able to provide that support to you.
Or you might just take these results and say, thank you so much.
Now we're going to go do what we think we need to do.
Going a little bit deeper.
We'll call this the maybe 8,000 foot view.
This example is broken up by department and location,
and it's showing how much time people
are spending doing data exploration and visualization.
We can see in the manufacturing department
and in the fermentation PD department,
a few people are spending over eight hours a week.
Whereas over at R&D, they're spending considerably less time.
Maybe that's fine, maybe that isn't fine.
But again, it just helps you see exactly how your people are spending their time.
If you have someone located in the east in the firm department
and they're spending zero time doing data exploration and visualization,
that might be a problem.
I would think they would need to share their data and look at that data.
So you might have some questions and you go ask that team
to better understand exactly how they're doing things.
Then we get to go a little bit deeper, and these are my favorite images.
What we're looking at is proficiency on the left
and usage on the bottom.
Y versus X.
Then these cells are colored
by how many people are spending time or grafted with their competency.
If we look on the left, we see quality process and engineering.
For this particular organization,
by and large, people are not doing much quality work.
This is where I always ask,
do you want your company to be quality minded,
or is quality focused on a single department?
If you want your people to be quality minded, this might be a red flag.
No one is an advanced user.
People are spending predominantly less than an hour a week,
and the majority of the people aren't doing it at all.
This would be a situation where I might come in and say,
can I teach your people about the quality tools in JMP
so they can better understand
how to build and interpret a control chart?
How they might look at metrics like CPK and PPK.
Are we hitting spec limits? Are we not hitting spec limits?
How do we understand that more deeply
so we can make more intelligent decisions to solve problems and understand things.
Contrasting that on the right
with this basic data analysis and modeling image,
we see, I would call, more maturity.
There are very few people that aren't performing this task at all.
The majority of the people are intermediate too with some advance,
and there's also a lot of beginners.
But what I see here is critical mass.
I see that this organization has enough competence
and enough people that understand it well enough
that they can help draw those novice users down to the intermediate level
and we have some advanced users
that can bring the intermediate down to their level.
We also have people spending, predominantly 1-4 hours.
Not that many people are spending 20 % of their week performing this task.
When I see that, I ask, could this be automated?
We'll get to more of that in just a moment.
This really helps people understand where they are.
Do they have maturity in this analytical capability
or do they need support?
Are they where they thought they would be
or do they need to move their people through support
to a more advanced understanding and competency?
This is probably one of my favorite images.
What we're looking at here is on the left on the Y axis,
is the amount of time people are using a particular capability per week.
On the X are those different capabilities
that we saw in the JMP analytical workflow.
For now, you can just ignore the color.
Those are color coded by the amount of time they've been using JMP.
That's one of the only JMP specific questions
in the entire survey.
But what we see here,
and I want to draw your attention right here,
is there are six people spending eight plus hours a week
performing data access.
If you have 1 person spending 8 hours a week,
that's 20 % of the week.
That means five dots equate to one annual salary.
Is one annual salary how you want to be spending…
Do you want to be spending that amount of time on data access?
Probably not.
I mean, it's not cheap to hire someone.
It's not cheap to support them, provide benefits and training,
and keep them motivated and keep them growing within the organization.
This is a very impactful image because it shows us where can we automate.
Clearly, you can see up in the very upper left,
that green dot, someone has already automated data access.
Whereas we're spending
1.2 annual salaries on data access,
and many of them are new users.
Half of them are only 1-3 years.
So could we come in and teach people about automation?
Could the person that has already automated this
sit down with these other six people
and teach them how they have automated their process?
Because if you can free up an entire annual salary,
think of what you can do.
I've worked with people that are in hiring.
I've worked with people that manage teams.
The common thread I hear is we need more people.
Either A, we don't have the budget,
or B, and right now in this environment, it's just sometimes hard to hire people.
If you can liberate an entire person
from a particular task, just think of what more you could do.
They could solve more problems,
they could help bring automation elsewhere.
They can automate data access for everybody potentially.
This is a really useful thing.
On the flip side, if we look at, say,
predictive modeling and machine learning, it's right here in the center,
we can see there are only two people spending any time at all:
one, one to four hours a week and one less than an hour a week.
We're spending a fraction of the time,
particularly compared to data access,
on predictive modeling and machine learning.
Perhaps this is not necessary in your organization.
Perhaps it's very necessary
if you are trying to understand
why aren't we hitting our manufacturing KPIs?
Why are we having these issues in our process?
We're not able to understand exactly why, despite having everything set up
the way we think should work, we're not hitting our metrics.
Well, again, this is a ripe opportunity for support.
Yet on the other flip side,
so I think we're looking at a triangle here, a prism here,
basic data analysis and modeling, I see they're doing fantastic.
They have a lot of people
that are performing basic data analysis and modeling.
We have, as we saw earlier, some good there.
We're not spending a lot of time.
These are one of the tasks where automation may not be possible.
It might be, but it might be people are
dealing with problems that are unique every single time.
Again, this is an opportunity where I,
as someone trying to support a customer, might be starting to ask some questions
and understand if automation is even possible.
But by and large, this is a good vertical in this particular graph,
as well as over on the far right, sharing and communicating results.
That's another one that is very easy to automate,
but the majority of people are sharing results.
I might have some questions
about why there's maybe about 40 % that aren't
and is that important to you?
But this just really puts the entire story
in one image that really helps you understand
where opportunity is to A, automate, B, train,
and C, say, great, we're doing well,
we don't need to spend time on that.
Then finally, where do your people think they need support?
On the left, we have those capabilities.
On the bottom, we have four questions.
Is this critical to my task and training is needed?
This is critical to my task and basic training is needed.
I don't need this, or maybe I'm just interested.
I've organized these based on the majority of people
or the number of people that feel that advanced training is needed.
This is from a different organization.
But I imagine if you are a tech-driven company,
an analytically- driven company, you would probably think
that basic data analysis and modeling
is not something that you need to worry about.
But here, the majority of the people
are saying this is critical to my task and I need advanced training.
Again, it's just shining a spotlight on the areas where you might need
to support your organization,
where you might need to support your people
because they're saying very clearly, I need help.
Whereas mass customization, automation and scripting,
reliability analysis, these aren't things that people need as much support on.
You can know that I don't need to spend time in this area or that area.
I need to focus up here.
It turns out a lot of these are fairly basic tasks
that I think a lot of people think
that people are fully capable and competent of,
but they're clearly saying, no, no, I need some help.
The benefits of the JAWS is,
it identifies strengths and areas for improvement within your organization.
You're able to work with your JMP support team
and provide support in those areas.
Your support team has the training, has the tools, has the backing support
of a large organization that is focused solely on expanding the use of JMP
to come in and guide your people in whatever support you need.
We are here to help you out.
Then the beauty of this is if you administer this survey annually,
you're going to start to be able to track progress.
You'll be able to see we needed a lot of help in design of experiment.
A year later, we see improvement.
The support has worked,
and we just need to go a little bit farther.
We can say DOE is now doing great.
Let's focus our attention elsewhere.
Some best practices.
This is very practical, but what we have learned is,
don't allow a long time for the survey to be filled out.
We say send it out on Monday,
send a reminder email on Wednesday, and close the survey out on Friday.
How this would work is you would work with your JMP support team
to craft those questions about region,
about department, about job title,
and then they'll just provide you the survey.,
You send that survey out and people fill it out,
we collect the data, it's all anonymous, and then we analyze those results
and then come back and share those results with you.
But keep it short.
Don't allow people a long time because people get busy and they just forget.
It's really important, I think, to get those three questions right.
We don't want to be too much of a grouper
because then you don't have
the level of understanding that you might want.
You want to be a splitter.
Really dig down, get those departments right,
get those job titles right, get those regions right,
because you can always group things together later
to understand the survey results,
but you can't split them once you've collected that data.
This one is probably the most important,
is it's incredibly valuable to get management buy-in
and then develop a team to help administer that survey.
You need people that people are going to listen to.
If you have someone in your company
who is, for lack of a better term, not well liked and they send out a survey,
people probably aren't going to be as likely to participate
as if you have a team of 3-5 people
that are leaders within the departments
or leaders within their organizations,
and people are going to hear that and listen to that.
It helps even more when you have management saying, you need to do this.
The flip side of this is when you get the data back,
you want management to be involved.
You want someone that has some decision- making capabilities
to see the results
and understand what's going on
so that they can help craft the big picture.
It's great when JMP usage expands from the bottom up,
but when you're trying to drive something at an organizational level,
you really need people that are higher up to help drive the usage from the top down.
We strongly encourage that you administer this organization wide.
Ignore JMP usage.
A lot of our companies have
email list group of strictly their JMP users.
But really, that's often just a snippet of the company.
We have found people and human resources
that when they see what JMP can do, like I've got to have that.
We don't always think of an analytical software tool as being something
that maybe HR would want or would gain benefit from, but they do.
By understanding where your entire organization is,
you're going to be able to make better decisions,
you're going to be able to make better support calls,
and you're going to be able to move your entire organization
versus just moving a single department or a single job title.
Lastly, I'm sorry, second to last, lean on your JMP support team.
They administer these surveys frequently.
They know how to interpret the results.
They know how to help you.
Even if you don't want to use them to provide that support,
if you have an internal education,
maybe you want to build internal education up
as a result of the survey,
but lean on them to help guide you because this is what we do.
We support companies like yours to help them build analytical excellence.
Lastly, and I've said this a few times,
administer the survey annually,
so that you can actually track your progress.
It's one thing to have an analytical snapshot.
It's a lot better to have an analytical time series.
Collecting that data annually
is going to really help you gage is this successful?
Do we need to change things?
Is what we're doing working.
Maybe you go to your support team
and they lead the training and you don't see growth.
So you turn to your internal education team,
or maybe the flip side is true.
We just want you to be better, we want to be collaborators with you.
We want to support you
in whatever you think is the best way to execute that plan.
I will close with a call to action [inaudible 00:24:26] survey.
It can really help an organization.
I've seen it help an organization.
I've been able to administer close to a dozen of these,
and all of them have resulted in...
I'm going to close with a call to action.
Connect with your JMP support team
and complete the JMP Analytical Workflow Survey.
Being able to understand where you are as an organization
versus where you want to be is incredibly valuable.
At the end of the day,
our goal is to help you democratize analytics,
help you have one version of the truth,
help you make analytically- driven decisions,
and from that, gain efficiency,
quicker discovery, and save money, save time.
This survey is an incredibly powerful tool
to help you achieve those ends.
We have the expertise to help you not only administer the survey,
but interpret the survey and create a plan
to then make decisions for training support,
and help drive you from where you are to where you want to be.
Thank you very much.
