Where is your Installed Base? How to get more visibility into your Installed Base Data
How to get more visibility into your Installed Base Data This session will cover: - Implementing Installed Base Visibility - Benefits of Installed Base Visibility - Enhancing Customer Relationships - Driving Revenue Growth
But Mr. Lullet is the CTO at Entitle and has been at the forefront of designing
both SAS and AI-based solutions in the industrial OEM space for decades now.
He's one of those industry leaders who loves to get down in the trenches and
has immense expertise in things related to the install base.
Lullet has a phenomenal grip on technology and its application for machinery
manufacturers and when he is not at his work desk, you will find him posting
some very interesting stuff around AI and data on LinkedIn.
So that Lullet's over to you.
Thanks, JW. Thanks for the introduction and good morning, good evening, good
afternoon to whichever part of the world you are.
So, yeah, so today we are going to touch on this topic of install base
visibility and install base as such is usually a very important topic for any
original equipment manufacturers, especially industrial OEMs who make these
large equipments.
And if we take a step back, everyone understands that inside that install base
if mind correctly lies cold. What it means is that basically if we can
understand the customer behavior, the equipment behavior that is sitting in the
market, it can give us a lot of more revenue opportunities.
Because once an equipment is sold, the relationship that it ends there, the
relationship continues in terms of the service needs, the parts needs, the
consumable needs and it even goes to further that when they want to replace
equipment, home they should go to and that is where the quality of the
relationship makes a lot of difference.
So, with that in mind, we will look into the full install base notion and how
to make sense of it, what we can do and how we can drive actually growth out of
it.
So, I am sure you might be seeing a diagram on the slide right now and this is
a very common story that's what we have seen and this is an endless hill for
most of the OEMs.
There are like so many systems and especially with large organizations with a
lot of mergers and those things going on and even with some of these companies
are like that we deal with, we have seen companies which are like 100 years old
kind of things. So, they have, they just don't have systems coming as part of mergers, but they
also have a lot of legacies systems in place.
We have even seen data sitting in mainframes and then what happens is that any
new thing is needed any new initiative sometimes will land up making new
systems or making new data silos.
So, essentially what it leads to, it leads to all kind of actually integration
message well, the data flowing from one system to another system and then each
of this data has their own peculiarities, they do become a cellar's with time.
And this all leads to very, very non-scalable way of doing things and it's not
unusual.
This story we have heard again and again that if they need to build even a
small part of the view of customer, lift customer 360 just an aspect of the
customer, it basically means a lot of manual work in terms of pulling the data
from couple of systems.
Actually, even before that figuring out where the data points are residing,
then basically pull it makes sense of it because somewhere the data that and
ontology is different, the categorization is different, then someone has to
make sense of it, put it together and then some analysis can be done on that
one.
The only problem is that sometimes takes so much time that by the time the rid
dles are there, they are not valid and why this happens.
This happens because as I have talked about like what happens is that lot of
these efforts go. Of course, like there are certain things we started
enterprise level like ERP choices, CRM choices and then sometimes there are
like those big divisions, everyone has their own way of, in fact everyone has
their own take on which ERP or which CRM.
Sometimes these are not driven in a from the headquarters basically. Again,
like no one model is right or wrong, but what happens is that these all leads
to so many different kind of decisions which fundamentally leads to so many
different kinds of tools and that's where leads us to this whole data mess.
So before we move ahead, let's just check the state of the things within the
ecosystem, with the crowd out here, all the participants who are there. So we
'll just run a small poll, we'll just try to figure out like what is the state
of the installed based data right now as part of your organization.
So if you can help me with starting the poll, you might see like poll on your
street.
So basically, they say that what is the state of the installed based data, are
you still mostly working on spreadsheets or the situation is that they are into
multiple ERP CRM systems or you have some sort of MDM or some sort of master
data management notion in place so that you can reach to one system and get the
complete 360 view.
And going further, there is a MDM, there is a single unified source and there
is also analytics layer on top of it so that more insights can be derived out
of it which can help in revenue growth.
Okay, so I think this is actually very interesting and not I would say not
different from what we have seen, we see 25% mostly on spreadsheets and 75% on
in multiple systems.
And this is what is the state of the thing right now, there is like it's what
we see day in and day out with our customer base as well.
So before we move ahead, this is what we have seen and this is what the sum of
the like art and all else have verified that what they see we have seen
typically 7 to 12 systems again like based on the size of the OEMs it can vary
a lot.
But we have seen generally 7 to 12 system kind of things and usually the
productivity loss because one of the fundamental reason is that someone has to
collect this data together makes sense of it, unify it before anything useful
can be done with that data.
And the other problem is that even if this unification happens, there is still
a lot of data quality issues that still lingers on and then to do anything you
have to reach out to more than one people.
So this all if you see take a step back what it all leads to it all leads to a
very, very highly un scalable system and the guarantee of it being accurate
being precise is also very, very low.
What's the solution and that's what we will get into we will try to figure out
that how that situation can be solved.
So what we want to do it what is the ideal situation the ideal situation is
that if we can have a single source of truth what it means is that a system of
records must MDM we can go it.
We can give it a different names but basically there is a single place to go
which is all the data unified together clean duplicate did duplicate it in
reach everything and it is available to all stakeholders and not just that
there are insights there analytics very AI layer on top of it which is helping
to create those predictive,
predictive insights and what happens is that when this data is centralized it
also helps in collaborations and what the end result and result is that we are
able to serve the customer the best possible way and going further this is
where the growth lies.
So and that's where we talk about this whole install based visibility and why
it is important and if you really look into that it's not just about one part
of the organization that gets help from it it's not just a sales team it's not
just the marketing team in fact if that data is clean we have seen in our
customer base people have used it to get inside right in pricing in inventory.
So all sorts of functions start getting help out of this clean data so the
question comes up how should we get to this picture the most usual way I think
again here I would say there's nothing different from what a typical if anyone
has little bit gone through this data cleaning exercise is not very typical
about what what someone has to do and this thing people do it what we have seen
is that to do some limited inside they do these things in a very very manual
way but again it cannot be done.
But again it cannot be done at the scale when it is done in manual way but
essentially what you do is that if you look on the left hand side of the things
you have a set of systems you take that all data together and a profile it to
understand the nature of the data and then you take it through a path of
cleaning it which would mean cleaning mapping enriching the duplicating and
then unifying and stitching it together so you can actually bring the whole
data into a unified data model.
And then once it is there the magic happens you now have a single source of so
this is a technical view of the thing if you look from the business perspective
what essentially is happening is that you are bringing all the data into place
here again like this is a typical thing if you go to any data unification
exercise the the philosophy remains seems but being a weird more focus on the
install base data but essentially what would happen is that you will start
getting a very very clean and unified picture of all your customer.
And then you have all your customer addresses equipments parts service context
wearing these order history service history whatever makes sense to understand
that customer whatever makes sense to when we say that this is our customer 360
and once that is in place then one whatever you do like about AI or all kind of
analytics that start becoming possible and because the data that is coming on
the left hand side is very very clean the quality of the ridges.
Automatically becomes much better they are accurate they are precise so that's
where and then they help in taking what you call a data driven decision and how
now typically this is done so again we have seen models like we have seen many
companies trying it the DIY way and what happens is that basically you on the
left again like is the left to right if you see you will again it will
essentially go to the same exercise of acquiring.
Data cleaning and then you run some analytics engine but in the DIY what
happens is that you have to now basically put a team different kind of tools
and then it takes a lot of resources and time and that's where this usually
becomes a very very costly exercise and it's again we have we actually it's
very interesting that we have seen in our customer base.
Some customers actually who has taken this approach and they did it for for a
year and year plus kind of thing and then they realize that this is not a very
scalable way of doing things and then we have worked with them and we are able
to do it at a very very high scale and why we are able to do it actually let me
tell small story here actually when the title started almost nine 10 years back
initially we thought that we will just focus on the AI part analytics part of
the things.
What we wanted to do is that we will get the data from our customers and we
have over all this focus which algorithm to find the patterns in the buying
behavior so that we can predict and prescribe that what kind of parts a certain
customer would be needing what is the propensity to buy and all those things
but what we realize is that the data we are getting is very very bad.
So what we decided on that unless we don't take control of the data story we
will never be able to have a very predictable story on the analytic side of the
things and that's where we started investing a lot on this part of the equation
as well and that's where we have our own like this installed with data studio.
What essentially it is that is that like again the same steps you profile you
clean the duplicate you basically then unify it together and then you get a
very clean set of installed based data and just putting some screenshot here
just to give an idea of the things this is basically a complete local is the no
code platform what we do is that we basically once we go through the initial
exercise we capture it as part of recipes and then whenever a new data sets
come in the data pipeline.
The data pipeline automatically run those recipes and this really helps us in
doing the data cleaning at scale and once that is done then what happens is
that this is available into our what we call this the front end so we have a
front end mobile and that one so I'll give a small view of that one that once
you are able to get through this exercise and you have a mechanism in place to
slice and dice this data it can really help in doing a lot of powerful.
Powerful data driven customer research let me just switch here a little bit I
'll go to so this is how once you have done that when if you see here this is
now all the installed base if you see here there are like equipment and
locations it is one of our demo account but you can now start going inside that
one and you can see that well all your customers are if I just click here it
can even tell me like this place how many equipment parts and all those things
are there.
And then what I can do is that now I can for example I want to go to location
call QNX communication and so we have a notion of account location what it
means is that so for example general motor general motors have 10 plans so
general motor would be an account and each plan would become a location so I
can go to a plus location out here and here if you can see that it is telling
me that at this location what is the behavior of that customer along the years.
And you see here it is telling me that there are 4 equipment so I can see all
the list of equipment in fact I can go to every equipment and can see the edges
bomb the service bombs sitting against this equipment and then there are again
like details what we talked about like we can see the parts list the services
service contracts was the status of the wear and t's context and this is
interesting thing opportunities.
So what we also do is that we basically look into this whole purchasing
patterns of the past on the history not of just this customer but across the
customer base and then we do some sort of there is like those models that we
have built on which we can figure out the behavior of the whole co-horge inside
the customer and we can then fair benchmark the best ones and there is an
opportunity where we know that the leg arts can be taken up to the level where
the benchmark customer is sick and then.
This is interesting that now if you really see what is happening here is that
the whole 360 view of that customer is available at your fingertips and you can
as well if I go back to the my map if you see here there's a lot of filters you
can really for example if I just want to know that show me the equipments which
are like 5 to 10 years old kind of thing so it is just filter and show me those
equipment so that's kind of what slice and dicing is possible.
So that you can really define your pipelines in terms of so for example if you
know that certain equipments of age let's say older than 10 years 10 to 15
years kind of thing can be upgraded or some sort of that thing those those all
things start becoming very very easy but having said that I think we always
hear that not another tools and that's where what we are doing is that we have
soon very launching.
Well okay let me take a step back actually this whole architecture is also very
open architecture we have all open APIs and everything it plays customers do
connect to our open APIs and take this data back through into their system but
what we have also going to do is that we are going to launch it at the sales
force package we know that like last of our industrial oeums that we work with.
Majority of our customer are actually in sales force what we have done is that
if you remember here if I go to the same QNX last location so we were able to
see the basically the 360 view of that location so if you see this 360 view
that this 360 view is now would be available right inside the sales force how
does it helps now imagine that a sales person is now researching an account and
they have to compete with the salesperson. And they have the complete history of that account available here in the click
of the button they can see what all equipment that customer has what all parts
services again the same thing that I just talked about it's exactly a replica
of the same thing but it is available right in the context of the workflow that
a sales person is running when he is doing or researching to find the
opportunities within that.
So yeah I guess that's what I wanted to cover today the whole notion of this
part of this visibility and how we can achieve it from the data cleaning part
of the thing taking it through the whole data pipeline bringing it to a point
that it can it is available to the not just sales team but across the
organization to take very very rich data driven.
Yeah, I think that's probably we can stop here and we can see that if there are
any questions we can take that yes awesome thank you all appreciate it we did
get a couple questions the first one you mentioned about a data mess in early
slides yeah what exactly are the services you off.
Okay so we are basically what we call this we have a stalvesth intelligence
platform so what we do is that we take care of this whole entire pipeline we
just so for our customers they just have to give us their data and then we take
care of everything we will take it will figure out of course like we need your
help in terms of understanding the domain because you people are the domain
expert you will tell us that like certain categorization or how the product has
been. How the product hierarchy should be made on what what are the classifications
but once basically that is in place we just take it through everything we run
it through our pipeline and we build it in such a way that it just remains
scalable next time things come up it just runs automatically the data is
claimed it is now available on insights is available in CRM it is available
open API you can integrate with your own systems and then we have analytics all
this algorithms built on top of that one that will start giving.
The opportunities the predictive opportunities we can we also tell about prop
ensity customer health and all those things what it does is that it basically
arms you with all the relevant data points to basically when you are talking to
a customer you know exactly what are the points that you should leverage to
basically when you were engaging.
Perfect thank you the next question if the wrong data was ingested initially
can we export those wrong information and update with correct information we
ingest absolutely this is our we do it on day to day basis and then we know
that sometimes it takes time to get it right and we have an iterative process
it's not like once and once done and those kind of things the data pipeline is
very flexible.
And and and we we always constantly look for opportunities to fix data we have
a lot of mechanism in place and in fact going further we also like we we work
with a notion of a very I would say that a little bit high touch time to think
so we have experts whom we call the customer success manager who understand the
after market very well who understood the data part of the story very well and
they were very constantly in terms of making sure that data is of clean and
very high quality perfect.
A couple more next one if we implement in title what is the implementation time
and learning curve for our team okay so to question the first part is
implementation time depends on the size of the things but typically we have
done like implementation within I would say 8 to 12 weeks kind of thing that's
where people start seeing their results on this day they are able to access
their data they can start even between pipelines size and twice the data.
And those kind of things to do with the second part of the question second part
is what was what is the learning curve for our team so if you see the UI is
pretty intuitive and it's not like we just leave you there is a very high touch
engagement that happens where our CSNs will make sure that your team
understands how to use the tool how to make pipelines there are constant touch
points. There are periodic meetings not just in the initial days but throughout the
engagement period as well and so that we make sure that you are leveraging the
tool the best possible way but otherwise the tool is very very intuitive if you
understand if I think any salesperson understand this whole customer 360
behavior if you just look from that perspective the tool is very easy to know
yeah. Awesome good okay one one more is your solution compatible with sales force why
do I need it if I already have sales.
Okay so the thing with sales force is that it doesn't have data it has the data
model but it doesn't have data and that's the piece that we think we bring all
data together so for our customers it's not unusual sometimes we break 10 20
years of data. So remember so so like one of our bigger customers is they have chillers the
chillers has life of 25 or something 30 years kind of thing so what we have
done is actually the whole 30 years of data here and we have cleaned it
together we now know the 30 year history of every customer location we know
that against that chillers what all has happened what are all the different
cellars chillers against that location this is where the power lies so sales
force and these tools they provide a good data model but someone has to fill
that data and that's where in title comes into picture we will take all your
data put it together unified
and we'll give you that data access so that you can take very high quality data
driven hopefully JW like yeah and such the question I think I think it
definitely does.
There was another one came in though you kind of just touched on it how does
entitled give visibility to our install based data is the UI tool and then if
yes is it only on top of sales force.
No in title in essence is the installed base data or the core of the title is
about building that data you can call it MDM or what sort of that thing but
essentially what we do is that we will take all the data in our platform will
take it through that whole data cleaning exercise we have we have built all
those tools that helps us in doing that in a very very scalable way that's
where like if you see our implementation cycle are usually like 8 to 12 weeks
kind of thing of course like if you can see that we have a lot of data in the 8 to 12 weeks kind of thing of course like if you are talking about very huge
install when I say huge it's like 100, 150,000 kind of locations kind of thing
and bringing their 30 year history but otherwise usually 8 to 12 weeks.
Customs are up and running with us we bring this data we will put it together
will give it the context we will enrich it we will deduplicate it and then that
is available through our inside
tool which I just show so I can just go back to that one yeah so if this tool
through which you can access it and this tool we don't have like per user
license kind of thing we just work on a subscription model where every one
possible in the
organization can access to that one that's the flexibility we give but we also
know that many of the sales organization wanted in the context of sales force
there are other part of the organization which needed in the context of their
tools.
So in those cases there are integration available the data is available through
open API so people make it available there but yes through this tool you can
now access this whole data people can come here research it or they can
directly see it in the context of their sales force.
So hopefully that answers the question.
I think so there was a follow up to that which is only master data or along
with ERP transfer data once again only says only master data or along with ERPs
Oh okay you can give us the RP data give us CRM data give your spreadsheets
give your mainframe data give your service FSM tools data.
We have we use there are some implementation we take data from 8 to 10 systems
and there are like we also take data from ticketing systems and all those kind
of things so give us all kind of data both data better for us because then
algorithms run at scale the richer the rich the data set is the quality of the
algorithm also becomes much better so not limited to any one system or anything
give us data from wherever you are to whatever quality it is.
And we will make sense of it we will unify it and we will bring it together and
we will put it in a canonical data model which is like the date the purpose
built in the data model for the install piece where we will host this data
perfect perfect all right we are coming just up on time folks well thank you so
much for the content and going through those those questions for the attendees
here.
Any closing remarks no I just want to say thanks thanks everyone to spending
their time last 30 minutes with us and seeing what we are doing and we hope to
talk to you some as well in the future.
Excellent yeah thank you for your time folks have a nice day or wherever you
are good I see you on the next event thank you.