Learn what’s new in Cognos from some of the people who know it best. On this episode, we welcome Michael Peter, IBM Technical Sales Leader for Analytics, and Chris McPherson, IBM Cognos Analytics Offering Manager, to share the latest updates on IBM’s business intelligence solution.
Jasmine Polkowske (00:15):
Coming live from Dallas, Texas is Jasmine from Cresco
International. You know what they say. Everything’s bigger in Texas and that
includes the stars we bring on our show today. Here we have with us Michael
Peter and Chris McPherson. On the Cresco side, we have Ketty Mobley joining us
to discuss what’s new in Cognos. I’m going to hand it off to Kenny now.
Kenny Mobley (00:36):
Hey Michael and Chris, thanks so much for being with us
today. But before we get started, I’d like to learn a little bit about both of
you and what you do for IBM.
Michael Peter (00:47):
Oh, sure. Thanks Kenny. So, I am part of what’s known as
the partner ecosystem. I work with a variety of our partners such as as Cresco
and am primarily focused on Cognos Analytics. My history is totally in Cognos.
I’ve gone back to 98 when I started with the Cognos and worked with the tool in
a variety of ways over the past 20 plus years. Today I’m mostly focused again
on working with our partners to help them understand better strategies of how
to get the product into their clients and make it successful.
Chris McPherson (01:20):
It’s Chris McPherson here. I’m the Offering Management
Leader for Cognos Analytics, so I lead a team of product managers who are
responsible for the Cognos Analytics portfolio, both on cloud and on premise.
And I’m based in Ottawa, which is where the Cognos R&D is centered. I’m
like a legacy cognitive person myself, like Michael. I was with Cognos since
2005 so I’ve been doing this awhile as well.
Kenny Mobley (01:46):
So it appears that in the last few years, Cognos has lost
some traction through rivals like Tableau and Click. What are some of the key
changes that are being made in an attempt to change this perception?
Michael Peter (01:58):
Well, you know, there’s no denying that we lost some
market share to those vendors and I think it’s important to understand why as a
context of where we are headed and the changes that we’re making. This market
has seen a pattern that shifts between control of information delivery and line
of business doing things on their own. You know, Cognos was built on
user-focused tools like Power Play but as those grew in an organization, it
needed more governance and scalability. So we evolved in that direction and
that made it difficult for the line of business users to get what they needed,
so they turned back to desktop tools and in that cycle it was dashboard and
data visualization tools like Tableau, you know, those are the ones that got
the attention. What’s interesting is we’re already seeing the next shift though
and I think a lot of that is due again to a lack of governance, but in
particular the lack of analytics governance. What I mean by that is
organizations are finding that having a tool to easily create a dashboard does
not always yield reliable answers
Michael Peter (03:00):
because not everyone understands how to do data analysis.
You know, our answer to that, and as we look at the future and where we’re
going to change that perception issue say is AI. And I think that’s why we are
working to make AI so pervasive in Cognos Analytics. Chris, you got some ideas
on that?
Chris McPherson (03:20):
Yeah. As you pointed out, Michael, you know, we really did
I think lose some market share just around the whole self service visualization
agenda, particularly on the desktop. You know, we’ve done a couple of things.
First thing we’ve done over the past few years is really worked on our self
service and visualization capabilities. So if you’re looking, especially with
11.1, in the past 12, 18 months, we’ve made some great progress in terms of the
quality and interactivity, dashboards and visualizations. So I think we’re now
at a point where we’re able to compete with those other vendors by providing
these capabilities that give you the same degree of functionality, but with the
governance that Michael talked about underneath. Even more importantly, we’ve
put a huge focus on AI and machine learning over the past two years. We’ve done
a lot of work building up the foundation and the infrastructure to be able to
support some pretty exciting AI driven features and in the past six to nine
months we’ve really been bringing some of these features to market. Things like
predictive forecasting as an example, the natural language conversation
assistant – these are just automated content creation. These are just a couple
of examples, but we see AI and machine learning as really an area where where
we can differentiate because you know, if you look across the market,
visualizations and dashboards have pretty well monetized. Everyone has pretty
pretty visualizations. Everyone has a dashboard. So we see AI in particular as
an area that we can really differentiate and create some white space between
IBM and some of our competitors.
Kenny Mobley (04:56):
Chris, I appreciate that and I want to dig in a little
deeper until the end to the machine learning and AI aspects of what you’re
talking about. But before that, I want to go back and talk a little bit more
about the self-serve analytics side. When I was at GameStop, we took heavy
advantage of the features within Cognos to move to a self service environment
to GameStop where it was all in the past, I have requests to it and us
delivering reporting, we were able to use not only the features that allowed
for easy report building and visualizations in Cognos, but also that governance
piece, which was very important when you’re trying to move that out into an
organization because it’s really difficult to just give people data without
some kind of governance on how they might use it. I want to ask a particular
question though about dashboarding. I know that a number of improvements were
made, but in your opinion, how does dashboarding differ from reporting? I hear
people use these terms synonymously, especially if a report happens to have a
chart on it. So how should I use or think about these things differently and
what has Cognos done to improve that dashboarding experience?
Chris McPherson (06:06):
So as I mentioned earlier, we’ve done a lot of work just
improving the overall quality of the visualizations, the degree of
interactivity. So I think really that’s the use case, the core use case for
dashboard. It’s for someone who wants to do data exploration. They need a drag
and drop. What you see is what you get a highly interactive experience, and
that’s what we’re providing in the dashboard. We’ve also built in some of the
embedded smarts, the Michael alluded to earlier,, but it’s really around a high
degree of interactivity and something that’s very quick and easy for a line of
business user or a new user to get started and start being able to get some
value. Reporting on the other hand is really designed for someone who has
specific requirements in terms of look and feel.
Chris McPherson (06:50):
You know, they may be financial reports but they really
need that pixel perfect degree of control. It’s also the answer for when you
don’t necessarily need to be live interacting in a very real time fashion with
your data. A lot of the use cases we see are professionally authored reports
or, or you know, authored reports that folks just scheduled to run overnight or
on some kind of a calendar with different prompt values to different audiences
and different formats. So I think both offering interfaces have a place and
we’re one of the few vendors out there that provide both capabilities out of
the box in a single product without anything else to buy or install. Michael,
do you have anything to add to that?
Michael Peter (07:36):
I actually kind of agree with both of you. Chris, you’re
absolutely right from the standpoint that the use cases of reporting as
traditionally defined and absolutely are accurate in that regard are different
from dashboarding and Kenny I agree with you as well from the standpoint that
there is a little bit of commonality if you look at the intrinsic use case of
both, they’re really about how do I get the information that I need. Right and
it’s really a matter of how easy and fast can I get the information I need to
answer the questions I’m trying to answer. No. For a long time reports were
really the only option for that and as the dashboard approach gain popularity
and the tools made it feasible, you know, that gain traction well as we talked
about before and it made it easier for users to find that information on their
own to a large degree. I think we’re continuing, as I said earlier, to see a
shift in terms of how that information is used. And the whole concept of self
service as you mentioned, comes into play.
Kenny Mobley (08:41):
Well let’s explore that a little deeper then Michael. What
to you does self service mean and how has self service evolved over the years?
As we’ve talked about using users having more capability to create stuff
themselves, but I get the feeling that self service analytics has a deeper
meaning than that. Could you explain it?
Michael Peter (09:00):
So it does in my mind because I think that as I was
talking about the evolution of Cognos, our focus is on very often been on self
service. How can that user find that information? You know, self service
reports were one thing in terms of creating those styles of reports on their
own. Same thing with dashboards and if you think about it from that
perspective, self service is kind of okay, how can I do something rather than
relying on an expert to do it? Mmm. And the same thing is true of self service
analytics. When you use that word is the analytics specifically as the context.
But it’s different in that when you think about the first two dashboards or
reports that expert usage was really, you know, how do I not write code might
be an easy way of putting it. It was a very functional type of, of expert. When
we get into analytics though, it’s not how do I do it, but more of the, why do
I do it? It’s how do I infer things? Why would I use certain sets of data? What
is the state of telling me what don’t. I know there’s so many facets to it that
really have nothing to do with how I created a dashboard as much as what is
this information telling you? And I think that’s the big gap that we are trying
to address. It’s finding the best insights, the information that’s really going
to make a difference to my business. Particularly things that I may never have
thought to ask. Those are the skills that when we talk about the expert in this
context, it’s the expert of the data scientist, if you will, in this world. So
how can we take that expertise that a data scientist or deep data analyst might
bring to the table and wrap that into a self service experience. Chris, talk a
little bit about some of the details. I know you’re deep on some of the
specifics we’ve done there.
Chris McPherson (10:54):
Yeah, I mean that’s a great explanation Michael. You know,
what I might add to it is, you know, in my mind, self service means how do I
get, how do I reduce the time to value for users. So if the users coming in to
have a question in mind or perhaps they just want to learn something new or get
some additional insight from their data, how do we enable them to do that? To
give them, as Michael said, the best insight and most reliable insight, but
also make it easy enough and fast enough that they don’t have to spend a lot of
time getting there. At the end of the day, that’s what they want to get the
answer and move on to the action and which is what this is really all about
enabling them to make the right decisions, smarter decision, and then take the
appropriate action. So I mean, you know, we’ve done so much in the past couple
of years around self service, particularly in the dashboard. You know, we have
literally started by picking a data source and then simply say generate
dashboard and you know, the product, we’ll build you a dashboard to use as a
starting point. So it’s these types of things that really remove the barriers
and just sort of reduce the entry point for line of business user or someone
without any kind of offering or coding experience to be able to identify
dataset, you know, whether it’s asking questions or dragging and dropping to do
data exploration and then get those insights quickly and be able to take
action.
Kenny Mobley (12:17):
Yeah, that’s great and it really leads us into the next
topic that we’re going to talk about and this is the natural language question
feature, which I think is kind of an evolved state of this time to value. I
mean, if I can just ask the computer a question and it’s going to respond to
me, that kind of cuts through everything else and get you to the point that
you’re talking about. Get you some insights and answers immediately. So talk to
me a little bit about this feature. What is it and what’s going on in the
background to make these answers feel natural?
Michael Peter (12:50):
Okay. So this is the conversation assistant that we
introduced on 11.1. It’s a natural language processing capability where you can
use natural language to questions to interrogate the data source and the tool
will provide a response so it really is a two way conversation. We’re specific
about how we refer to it as conversation assistance because unlike some other
solutions out there, you know, it’s not just a simple, you ask a question, you
get an answer and that’s it. Uh, it really is a two way communication where you
may get a clarifying question, you may get a visualization as a response. You
may get insights that you didn’t ask for and that you may want to look at that.
So it really was designed to be a two way type conversation between the user
and the solution and that it does make it feel more natural. It’s not just
simply typing in words and getting words in response. What we have is a an
ontology behind the scenes. This is juts a collection of terms that is
constantly evolving, but it’s a collection of terms that are relevant across
all industries. So as you’re typing things in it’s recognizing some of these
concepts and it’s making decisions on what to show you and things to identify
in your data. It’s running all of us in the background and then providing you
with the best. That’s fine. Awesome. We’ll answer. You know, the neat thing
about this is that we designed it in such a way that you’d be highly
extensible. So we certainly envision, you know, in the not too distant future
allowing a company in a particular industry or particular vertical to be able
to bring their own ontology so that they have industry specific terms or
perhaps, you know, in a large company they might have very specific ways of
referring to things and they could bring their own ontology or modify our
ontology so that their users get to the best possible outcome.
Michael Peter (14:44):
In fact, we’ve had another group at IBM do just that.
They’ve brought their own ontology and have some help from us wired it in so
that they are able to ask very specific types of questions and get very
specific types of responses based on these, the dictionary of terms that they
use on a day to day basis. So it’s pretty cool technology.
Michael Peter (15:04):
And Chris, you know, the thing for me, when we talk about
the natural language being in a dialogue, it’s so much more than just two way.
It’s all the natural language context that Cognos is adding to the picture now,
particularly something like you get into our explore experience, which I love
because it’s purpose built for this type of interactivity, right? Versus just
creating things on my dashboard. I mentioned that earlier, that, you know, dashboards,
sort of that visualization aspect, but not necessarily analytics. You know, it,
it gives you all this natural language context about what you’re looking at and
things you didn’t think of. One of the things I love often seeing is how it
will put the chart on the screen says that you can see visually that something
is, is higher than the other things on the screen. It will add a simple word
like unless or unexpectedly higher. Then that’s indicating those things
happening behind the scenes statistically to show that this, this is
meaningful. This is not just simply higher, but there’s a bigger difference
there then you might expect. It’s that elimination of, of uncertainty, that
elimination of bias that I bring to the table. Um, that, that happens by giving
me all this additional information and natural language that I don’t have to
think the data science is out of the gate. It’s guiding me through that
process.
Chris McPherson (16:24):
Yeah. It’s interesting you say that Michael, because you
know, I’ve had various conversations with customers. So you know, over the past
year or so since we’ve introduced explore and it’s amazing to hear how many
people actually, their eyes are just drawn to that text. You know, you’ve got
this nice visualization in the middle of the screen, but it’s the text that
they’re looking at and the reading to get these insights. So yeah, it’s
interesting how the combination of having those visualizations with the
accompanying texts just drives home its more engaging. It just really lets you
absorb or consume the insights that we’re providing in a much more effective
way.
Kenny Mobley (17:01):
I agree as well as I’ve seen that it is fascinating to see
it draw insights because of course the reasons that we’re looking at the
visualizations is to look at them for insights that visualizations could show
that perhaps data tables could not. So it’s very interesting to me to see how
y’all have been able to embed that within the cognitive system in order to help
users. Which leads me right into the next piece I want to talk about when it
comes to Cognos and its advanced features on the machine learning and AI side.
That’s something that I’ve heard referred to and I know IBM refers to often as
embedded smarts and this is the feature that learns and adapts to a user’s work
style. So it’s a different way of using this. But I’m guessing it’s somewhat
similar, but what does this mean to the user, this embedded smarts?
Michael Peter (17:47):
So when we talk about smarts and embedded smarts, it is a
fairly all encompassing term. It’s sort of our simple way of referring to a
variety of the technologies that exist in inside the product to do all the
things that we’ve been talking about so far. At its core is, is the concept of
AI, right? It’s important to recognize AI. I think the way we look at it, you
also have to dissect that a little bit. You know, when people hear the term AI,
the immediate audio is artificial intelligence and that deals with, you know,
the deeper data science and machine learning and deep learning and things of
that nature. But also AI, it has the context of what we call augmented
intelligence and that is very user focused. That is how do we use the AI, the
former description, the machine learning and things like that to augment the
intelligence of the user.
Michael Peter (18:46):
How can we do things easier for them? How can we do things
smarter? How do we, as I mentioned, eliminate bias any, anywhere and everywhere
that we can take something that a user is trying to do and make it easier for
them. That’s how we can apply the quote smarts. You know, I think that as you
mentioned, part of it is learning user approaches to things. A lot of it is
training we built in already. Let’s take for example data visualization, right?
There’s lots of tools out there that do them. We do them as well. But what’s
interesting is how much time we spent in terms of training the tool to look at
the data in a fairly deep way and understand, okay, with the data this user’s
requesting, what is really the best way to visualize that?
Michael Peter (19:36):
And then it will render the chart appropriately. But we
recognize that there’s lots of preferences, lots of charts that options that
are out there. So the tool starts learning. If I constantly reflect the
different one, I shouldn’t have to do that every time. So now the tool will
actually say, okay, instead of a bar chart, I’m going to give you a scatter
plot or whatever the case may be. You know, because that’s what you tend to
prefer and it will constantly adapt to that. So the smarts is all the stuff
we’ve programmed in and designed into it along with continuous learning that
we’re trying to get it to do to address the user work styles that you
described.
Chris McPherson (20:15):
Yeah, that’s a great example, Michael. The visualization
recommender, you know, we’re evaluating the data that you selected for gesture
basis, basically to show you the correct visualization at the right time. You
know, we do that. That’s kind of a follow on effect of something we do a little
earlier in the process where you select your data source, we identify the kind
of data types that you’re using and we can determine what’s a measure, what’s a
category, what’s a geographic location, what represents time so that we use
that information when you’re dragging these things onto the canvas and make
those smarter decisions to give you the best possible visualization. You
mentioned augmented intelligence. I really do love that term because I think
what we’re doing is really is just that, you know, you’re building, you’re
dragging and dropping or you’re looking at a visualization somebody else built
for you.
Chris McPherson (21:07):
And you know, without even asking it, the product is
analyzing the underlying data, looking for patterns, looking for outliers and
predictive drivers and then giving them access, giving you access to those things
really right, individualization with the click of a button. So we made some
really great and really exciting progress on that topic. In the last couple of
releases, most specifically with the time series forecasting that we brought
out in 11.1.4 where literally with a single click will, the product will run a
number of different algorithms independently to pick the right one so that you
get a statistically accurate forecast right in place. That’s just one example
of what Michael mentioned, augmented intelligence where we’re not replacing the
human mind, we’re just showing them insights to make their job easier and allow
them to make smarter decisions to take better action.
Kenny Mobley (22:04):
Yeah, that’s great and those are some of my favorite
features. Those of us that have been in analytics for a long time know that
choosing the right visualization and doing the things you need to do to make it
look good and be correct can be difficult and take a lot of work and having
some kind of augmented help definitely is something that’s valued. I’m glad
that y’all are putting the time and effort into that. I want to talk about
something a little bit more practical now and that’s about getting into using
Cognos. I think there’s some idea out there that using Cognos as a large enterprise
tool is very difficult, but is that still the case? How easy is it to get into
Cognos these days?
Michael Peter (22:48):
Yeah, it’s funny because we started the conversation
around perception, right? And I think that is still unfortunately a perception
that people have to get started with the Cognos. You got to get it involved and
you know, big implementation, lots of time and effort and it really doesn’t
exist anymore. Literally today. Anyone can go to ibm.com, or Kenny your website
and, and get started with a trial, download it, start working with it. It’s
completely cloud based. You can get access to your on-prem data. All of that is
available to you to get started right away. Or you can start with a spreadsheet
and start analyzing your data from that way. This nice thing is the trial is
free for 30 days, but then you can go on and immediately just transfer that
into a single user license if you want for as little as $15 a month, you can
start working with Cognos and leverage, know much of the power that we’ve about
today. If you want all of it, it’s a little bit more, but still it’s something
delivered on the cloud and very easy to get started with. And of course, you
know, as companies grow, if they outgrow what we can do on the cloud, companies
such as Cresco can, they can certainly help migrate into the larger
environments as well. But we’re finding so many more people are able to get
started now, just to get access to the tool now, once you’re in it, even the
whole, the idea of how do I use it, we’ve addressed that with, it’s a
tremendous amount of online or should I say inline tool, so right within the
interface, access to videos and tutorials and examples, just a variety of
things that help you get over the common humps that we see with users. In fact,
we’ve done a tremendous amount of study watching users use the tool and it’s pinpointing
those spots where they tend to have issue and we’ve addressed that. So both in
the design itself, but also in these tutorials. We’ve expanded our community on
ibm.com so lots of interactivity with users asking questions, sharing
knowledge. In fact, very soon we’re gonna introduce what we call an accelerator
catalog, which takes that concept of samples to the next level where there’s
going to be extensive array of different types of content that you’ll be able
to walk through and use and understand exactly how to solve a problem. Very
similar to your own right so very more tailored content to allow users to get
access again the types of things they need to get started.
Kenny Mobley (25:16):
Yeah. Yeah. That’s perfect. I really look forward to that.
I love these accelerators. Fast starts things that help people, you know, up
and running in a bench. A lot of people will learn now that it is easy to start
using Cognos in their, in their own environment and it being available on the
cloud really makes it not any problem at all to get started and start to take
advantage of some of these great features. So now that we’ve talked some about
what Cognos is, what it does, how it might be used, what are some of its
benefits, I’ve really liked that. Chris, what’s next for Cognos? What things do
you all have on your plans for the next couple of years?
Chris McPherson (25:53):
In the very near term, we have a new release coming out
probably next week, 11.1.6 which is going to have some great new feature- some
enhancements to conversation assistance, some enhancements to reporting,
dashboarding, visualization. So lots of great stuff coming in the 11.1.6
release mext week, but on a larger scale or more strategic area, we have a
handful of areas that we really want to focus on in the coming year to 18
months. So first, we talked a lot about AI and that’s going to continue to be a
big focus for us. That means both developing our own capabilities but also
leveraging other technologies coming out of Watson APIs, IBM research and
elsewhere and IBM. So we are going to be investing heavily in AI and just to
take those capabilities that we’ve got to the next level.
Chris McPherson (26:47):
The second area is the integration we have with planning
analytics. So, you know, we do have a number of joint customers, customers who
use Cognos Analytics for the reporting and analysis. They use planning for the
budgeting forecasting. These two products just have so much synergy and there’s
such an opportunity there to really get them more tightly integrated so that
you can really move from planning and budgeting into tracking and measuring how
your plan is executing and then feeding those up which comes back into your
planning process. So we’re going to spend a lot of time and effort bringing
those two products closer together, making that integration much, much tighter.
And we’re also doubling down on our cloud investment. So, you know, Michael
talked about our on-demand cloud offering a few minutes ago, you know, we’re
going to continue to build up the capabilities there because I think we really
do have one of the most flexible deployment options in the market as he pointed
out. You have a on demand cloud, which you can start with a single user. For
example. We have our dedicated cloud on premise of course, and we have hybrid
cloud offering. So like cloud pack for data, Cognos analytics for example. So
we really do want to invest in that area and really be cloud first so that we have
the right solution at the right time for our customers who are looking to
modernize or move their workloads to cloud because we know that many of them
are thinking about that. If they are, if they’re not moving now they’re, you
know, somewhere on their horizon is a move to cloud, whether it be for cost
savings, whether it be for modernization, we want to make sure we have the
right tools at the right time to help them on their journey to cloud.
Kenny Mobley (28:29):
That’s great. Thanks Chris. I really appreciate all that
insight y’all have been able to give us today on Cognos and let us look a
little deeper, not only into some of the functions that Cognos provides, but
some of the things that that team has been thinking about and how y’all make
the product better for everybody and some of the thoughts that goes into that.
So I appreciate all that.
Jasmine Polkowske (28:48):
Thank you so much Michael and Chris for taking the time to
talk with us today. It was great to hear your thoughts on Cognos and thank you
to our listeners for tuning in today. Please be sure to follow us on LinkedIn
under Cresco International and on Twitter @Cresco_Intl to stay up to date on the
latest news in the analytics world.
