Latest / The DEX Show: A Show for IT Change Makers / Reality Bytes #31: The Encyclopedia of DEX (w/ Guillaume Charles)

Transcript

Reality Bytes #31: The Encyclopedia of DEX (w/ Guillaume Charles) · The DEX Show: A Show for IT Change Makers

0:01
Welcome it, change Makers to the Deck Show with Tim Flower and Tom McGraw.
0:07
Let's get into it. They would change me because Welcome to the Deck
0:12
shows show within a show within a show. It is one of the thirst
0:15
special episodes of Reality Bites where various members of the Wonderful Reality Bites team get
0:22
to interview a wonderful colleague and Next thinker we have today. We've got Sean,
0:28
Tim and Oriana here with us and we are going to be talking to
0:32
the fantastic Guion Shall. He's product management director for diagnostics here at Next Think.
0:39
Gion. Welcome to the show. It's lovely to have you here.
0:42
Thank you them and fantastic and you're you're you're one of a series of kind
0:46
of more technical product based guests and we've really really loved having you all on.
0:51
So we're really excited to talk to you today. But I wonder if
0:54
you could just start by telling us a bit, a little bit about yourself,
0:57
and first of all, you're rude to Next Think and how you came
1:00
to join us here a few years ago, as I understand it, Yeah,
1:03
definitely, so I John six years ago as an individual contributor product manager.
1:10
As you can imagine the company was quite different then. It was much
1:12
smaller, the product was complete on prem engineering and product teams were very small.
1:19
Yeah, so I worked on lots of different features for the B six
1:23
products. Then I was the product manager for the first product that we built
1:27
on the cloud, which was and still his application experience. And now I'm
1:32
a product director managing five product managers. I'm based in Lausanne. I'm originally
1:40
from friends, well a few other companiers proct matgers before, and so that's
1:47
me now. Quite quite a successful six years you've had here. You have
1:52
to give us all some tips for success. We could do with them here
1:53
in reality, but it so, but give us give us a snapshot of
1:59
what you of, what your day to day responsibilities look like at the moment,
2:02
and the kind of projects you're managing right now. Yeah. So,
2:07
as I said, I'm managing a team of five very talented product managers.
2:10
So this in itself is already taking a lot of my time and curing my
2:16
role is ready to work on breaking silos, making sure that the different products
2:22
that I oversee and that those product managers are managing, you know, work
2:27
well together, that we build a common vision, making sure that we solve
2:30
the right problem for customers, that we build the right thing. So there's
2:35
a lot of alignment, lots of talking with you know, internally between ourselves,
2:40
but also with people from the fields, with our sales people, with
2:45
our technical services, and of course with our customers. And I'm also trying
2:50
to take time to evangelize our product and our vision. A fellow evangelist,
2:57
I love it. I'm really glad you're here. It's a lot of fun
3:00
to have techs on and to have a techi that's also a next manker.
3:05
I think it's pretty cool talk a little bit about how a new product,
3:13
new software product, that next thing comes about. It's obviously it starts by
3:15
developing something that the market needs. How do our product development teams get those
3:22
insights into the needs of the IT teams? How do you know what's needed
3:27
and what will have enough value for people to actually buy. So it's quite
3:31
complex. It's really the core of the role of product management, to be
3:37
honest. It's more not than the science. And there are many different channels.
3:40
So of course we're talking to customers as much as possible. That's very
3:46
important. We're trying to even visit them on site when we can, but
3:52
we also talk to people from the field, so to the sales reps,
3:54
to the CSMs, the solution consultants, professional services and so on and so
4:00
forth, because those people are working daily with customers and they're in a very
4:03
good position to know the various challenges that our customers are facing and the problems
4:10
that they have. We also have telemetry about the products that are already released,
4:15
so we can know which features are used a lot, which ones are
4:18
not, so it gives us some quantitative data. We have a portal for
4:24
customers to post their ideas, which is very useful for us, and they
4:29
can also vote for other ideas. We have a research team in the product
4:33
organization. By the way, O and I is part of this team and
4:39
they're super helpful also to help us, you know, do some large research
4:43
at scale better understand some of the specific problems that our customers have. And
4:48
finally, we are spending a lot of time looking at things like you know,
4:53
analysts, market trends, what the competition is doing, how new technologies
4:58
can unable to sort of some of the programs that might have been difficult to
5:01
serve before. So of course these days we are talking a lot about generative
5:05
as you can imagine, which opens a lot of possibilities. That's interesting.
5:14
What are you I'm just envisioning a list of things, right. The feedback
5:19
gets you to a place where you can start to list out and prioritize.
5:25
What are some of the long standing or running challenges that are faced by it.
5:30
How do you and how do those influence your list and what you focus
5:34
on? How do you take that list of inputs and prioritize and what are
5:40
some of the key things that are struggled are struggling with today. So there's
5:45
definitely a lots of things which have changed in the last few years, as
5:48
you can imagine, since the pandemic, the rise of looking applications becoming SaaS,
5:58
the adoption of twos like ends, and also the rutuous network access.
6:01
So all those things change the way it works, and now they face new
6:06
challenges. So more and more things are out of their control. You know,
6:11
they used to own a lot the infrastructure, the networks, a lot
6:16
of the things around them. You know, they had access to it.
6:18
They could go to a server and reboot it if needed. Now it's not
6:21
the case anymore. So there's very an increasing complexity, and even if it's
6:28
not in their hand, it's still very difficult for them because when something stops
6:32
working, the blame game goes on. It's very difficult to know if it's
6:35
something on the vendor side, on their side, on you even on the
6:39
internet, on the Wi Fi connectivity of the person working from home, for
6:44
instance. So you know, I imagine an issue as common and as simple
6:47
as someone using Salesforce or SEP and you know, suddenly it gets slow or
6:54
they're having an error, they cannot load the pitch. The issue can be
6:57
anywhere on the user's device, on chrow, on the Internet, on the
7:00
WiFi. Those applications can also be heavily customized, so it could be you
7:06
know, customization done by the organization which goes wrong, or it could be
7:12
the application itself on the vendor side. It could even be the data centers
7:15
where the application is hosted, like AWS or aisure. So it's a lot
7:19
of things, lots of changes for it, and they have to make sense
7:25
of all of these. They you know, they have lots of tools,
7:28
but they might say things which are contractory. They don't have the vision from
7:31
the end user, from the consumption like Nextink is providing. So you know,
7:35
we have to take into account all those challenges and based on all the
7:42
you know, the inputs that we saw that we discussed a few minutes ago,
7:46
and also of course the strategy of the company of Nextink and the vision
7:50
where we want to go. That's how we will prioritize our efforts and build
7:56
our vision. I know you and I have talked a lot about sort of
8:03
how so many of these things, like you said, are very superficially simple,
8:09
but under the hood they're super complex. And the next thing we talk
8:13
so much about see diagnosed fix. I'm wondering if you can talk a little
8:16
bit more about how that looks from an individual standpoint, right, if somebody's
8:22
using our platform, what does that look like in practice to try and untangle
8:26
this massive amount of complexity. Yeah, definitely, So three simportant words see
8:33
diagnos fix. But for each of them we could spend howls discussing about them.
8:37
So I'll try to transform those howls in a few minutes. As you
8:43
can imagine, they will depend on each other. Also, the first step
8:45
is c you know, we often hear customers thing that when they get next
8:48
things it's like you know, having the light zone, switching the light zone
8:52
and they start seeing things that they could have never imagine. So to be
8:56
able to see we need to correct the write data, so which means data
9:01
from the device, you know, device performance. I would say the most
9:05
the classics stuffs, the stuff we've been collecting from from the one like CPU,
9:11
memory, GPU and so and so force. But now, as we
9:16
said also before, more and more things are moving, you know, towards
9:20
the web, towards as applications, business applications. Also collaboration tools have become
9:24
extremely important since the pandemic and the rise of hybrid work. So we need
9:30
also to collect data to have visibility on this. So findstance to be able
9:35
to know what happens when people use web applications to those important business applications like
9:41
the salesforce to share point, the SAPs, you know, is it slow,
9:46
is it performing well? So we need to get this visibility that was
9:50
probably less important maybe five or ten years ago. And same for the collaboration
9:56
tools, you know, house the call quality. So that's the data we
10:00
need to collect. Now we need to be able to visualize this data in
10:05
some way, So we need to be able to trigger alerts relevant alerts.
10:09
We need to have dashboards, data visualizations, so all this is like the
10:13
part to be able to see. Then there is diagnosed, which is another
10:20
piece of work. There's a lot of data. Probably IT teams and other
10:28
computing teams already had a lot of data even before it starts using next Thing,
10:31
because they have other tools a PMS and PMS, you name it.
10:35
And the goal is not just to show data at them and you hope that
10:39
they will figure something out. So we need to be able to give them
10:41
insights, recommendation, and to help them diagnose. So we need to explain
10:46
why an alet was triggled. We need to be able to help them find
10:50
the root cause. And we won't always be able to find a root cause,
10:54
that's for sure, because very often the root cause is not even something
10:58
that we can detect. You if it's an issue with the application back end,
11:01
it's not something next thing can detect, and it's not something that the
11:05
IT team can detect or fix, but we can help them. We can
11:09
exclude other causes. We can say it's not on the device, or it's
11:13
not on the network, so it's probably on the application and we can avoid
11:16
the blame game, and that's something we hear about this meantime, to innocence
11:20
is super important. Being able to say it's not me and be able to
11:24
go to the vendor and confidently say you know there is something wrong on your
11:28
side, please fix it now. Diagnostics you can do this, you know,
11:33
device by device. This is very important. For the health desk.
11:37
We have to use Amplify or the device view for this. But where we
11:41
put more and more focus is also to diagnose that scale, because that's super
11:46
important. It's much more as you can imagine, it's much more impactful when
11:48
you can find an issue affecting ten thousand devices and be able to diagnose it
11:56
than to be able to go through all these devices one by one. Of
12:00
course, usually after a few devices you start to see patterns, but still
12:03
it's a tedious, manual process. So that's where we put most of our
12:07
efforts. And for this we have something which is really important and differentiators from
12:15
others, is that we have access to a huge data sets because our technology
12:20
covers millions of endpoints and therefore we can get insights that would be otherwise impossible
12:26
to get. For instance, we can see if a specific version of a
12:31
binary combined with a specific version of an operating system is behaving poorly, is
12:39
causing high CPU or more crashes than you know, the average of other versions
12:43
and ose versions. And that's something very powerful because first we can find insights
12:50
that we would not be able to find just looking at one specific customer data.
12:54
But it means also we can even inform customers when they're just starting to
12:58
deploy a new version and we can tell them, hey, be careful,
13:03
this version is bad, or even before they started plowing the version. And
13:07
then that's how you become really proactive. And that's really because we have access
13:11
to this huge data set and because of course we are leveraging new technologies,
13:16
powerful technologies to analyze this data set. And with this approach, organizations can
13:22
start you know, fixing issues at scale and hold vendors accountable. And that's
13:28
something that we hear more and more, which is becoming super important because as
13:31
we said, already, more and more things are now out of the control
13:35
of it and therefore they need to be able to talk to the vendors and
13:39
to hold them accountable. So that's a diagnose and to finish with the fix
13:45
I'll be a bit shorter. So at texting we are the pioneers with nexting
13:50
acts, so capability to execute scripts on the endpoint to remediate I'm proud to
13:56
say that today, a few years after we don't act, we have billions
14:01
of remeditions executed across all our customers every month. So you can imagine the
14:09
amount of time and tickets that he saved thanks to do all those automations.
14:13
And there is more to come. So stay tuned, attend our Experience Everywhere
14:20
event because you will see what is the next level in terms of fixings fixing
14:26
things. So that's the diagnosed and fixing. I hope it was not too
14:31
long, but there's so much to say about this, and you would know.
14:35
I feel like you have probably the most sort of comprehensive view of a
14:39
lot of this. You're sort of I think of you as our encyclopedia of
14:43
dex in a lot of ways, because, like you were saying, you
14:48
talked to so many people all the time about this. But I'm also curious
14:52
about sort of the rest of tech space. Right you're talking a little bit
14:56
about what makes us unique, but I'm curious if this sort of see diagnosed
15:01
fixed philosophy is adopted by other organizations. Do our peers and competitors generally take
15:09
this approach? Do they take a different approach? Well, I think everyone
15:13
is trying to see, diagnose and fix because in the end, that's you
15:16
know, that's very important. It's a huge part of the value proposition of
15:20
the DEXT tool and it's victually a loss cost saving and to improve the experience
15:26
of empoys around the world. Now, the approach taken to achieve this can
15:33
be different for each each organization, each vendor, and as of today,
15:39
most of our If I talk about the competitors, for instance, they are
15:43
not using the same approach for various reasons. For instance, they don't have
15:48
access to the same amount of data that we have and therefore their approach is
15:54
still I would say, to correct a lot of data as many it appoints
15:58
us possible with very high granularity and give this too and user computing teams,
16:06
hoping that they will have some experts, very strong experts. You note that
16:14
help three support engineers. Sometimes even some organizations have help four support engineers,
16:18
so they are really the top experts in terms of device truggle shooting, and
16:22
those people need to make sense of those data. So of course it's it's
16:27
very powerful. I imagine your trouble shoot one device in real time. We're
16:32
talking about sometimes you know, granularity of a few seconds. You have a
16:37
lot of data at your disposal, but as you can imagine, it's not
16:40
really scalable neither. So it's very difficult to trouble shoot that scale to see
16:44
the type of trends that next thing can see with all the data that we
16:47
have access to, and it requires very deep knowledge about hardware, network and
16:53
so on and so forth. Not so many people have this level of knowledge,
16:56
and definitely, you know, it becomes more and more difficult to find
17:02
people who are dis level of knowledge because just the complexity of the ecosystem is
17:07
increasing every year with new technologies, new tools. So I think we've showed
17:14
then the right approach at next thing. Of course, maybe bias probably,
17:18
but when I look outside, I think, you know, other tools are
17:21
not there yet. Maybe they will get there at some points, but for
17:25
now they're more on this, like, yeah, let's get all the data
17:29
possible and see what happens idea. I think so you sort of answered this
17:34
question already in the previous one. But so if you have to repeat a
17:38
few things, that's okay. But I was wondering if you could maybe just
17:41
expand upon a couple of examples of what cannot be done by another vendor who
17:47
purports to doing the same approaches the next thing, Like, what do they
17:49
typically face the problems they encounter? Yeah, so maybe what I could give
17:56
is a few examples that I've you know, I've heard of recently with some
18:00
of our customers, some success stories based on this technology, this approach that
18:06
we have at Nexting, and that other tools would not have been able to
18:10
vesically to catch. For instance, we have a customer who started deploying a
18:14
new agent for the taras and suddenly they realized very quickly that to disphace deployments
18:22
of not depoying of course on all devices at once. So each time they
18:26
were deploying a new ring, people started complaining that their deviceing was becoming slower,
18:32
the battery was battery life was shorter, and so on and so forth.
18:36
So I think they didn't really need even next Thing to realize that it
18:38
was linked to the deployment of the agent, because it was obvious. But
18:41
to our Nexting helped is in a matter of minutes, they could go to
18:45
a binary profiling and see that it was a specific version of this DLP agent
18:49
which was very bad and causing high CPU and impacting the devices, and they
18:53
could see that there was another version that has just been released by the vendor
18:57
which was much more stable, and then simply they could switch to this version
19:02
and solve the issue. So that's one example. Another one is we had
19:08
another customer who has a call center and they were using a specific application to
19:18
the cause and it had been a few weeks that when the employees were having
19:23
we're in the cause, we're having the cause. Everything was getting slow,
19:27
so the core quality was bad and you can imagine the call center, so
19:30
they were in touch with customers, so it was really having an impact on
19:34
the overall organization, on the business, and they could not figure out what
19:38
was happening. It was like nobody knew. And then with next thing,
19:42
we saw once again, you know, with this cross customer data, we
19:45
could see very quickly that there was once again a specific version of this tool
19:51
which was causing a high CPU during the cause, and the high CPU very
19:56
often, you know, is impacting the user experience. Of things be comes
20:00
slow and sole and so of us. And I can tell you it was
20:02
impacting tens of thousands of employees in this school center, so it was huge.
20:07
When they discovered this, they were like, you know, they knew
20:10
something was wrong, but they could not asss the impact and especially they could
20:15
not find the root cause. So this helped the rot. The last one
20:19
is another customer, so they have a security agent, and with next Thing,
20:23
they could see very quickly that on some devices this agent was consuming constantly
20:30
two gigs of RUM, two gigs of memory. So it was not like
20:33
this one was not impacting thousands of devices, you know, it was a
20:37
few, but still when you deploy a security agent, the last thing you
20:41
want is the agent to impact the device and to take two gigs of memory.
20:45
And hey, we could see not only that was the case for this
20:48
organization, but with next Thing and a cross customer data, we could see
20:51
easily that the distribution was the same for all customers. So you know,
20:56
if you look at the average memory consumption, it's okay, but you look
21:00
at the distribution, you see that there's always a long tale of devices which
21:03
are suffering from this, and for all organizations having this security a jump on
21:07
this specific version. So once again it was very useful. In these three
21:12
scenarios, the three different organizations open the ticket, discussed with the vendor,
21:19
and the vendor acknowledged the issue. And each time what they told us when
21:23
we discussed with them is that the vendors take them much more seriously now when
21:29
they come up with facts with nexting with data, and also even more when
21:33
they say that other organizations are impacted and they can prove that impact. It's
21:37
not only them, but it's impacting hundreds of other organizations across the globe.
21:41
And that's the type of things that only next Thing can find excellent. Thanks
21:48
for those are very clear use cases. You've hinted at this previously. I
21:53
was wondering if you can talk a little bit more about it. But I'm
21:56
curious of any of the implications that might occur for businesses who do have a
22:02
similar next Think approach of se diagnosed fix that type of detailed, thorough approach.
22:08
Do you think there are any implications that come with that, like,
22:11
for example, I know you mentioned L three support. They might, for
22:15
example, I'm imagining, find that some of their works has become redundant because
22:19
the technology is able to kind of do the heavy lifting for them. Are
22:25
there any other cases or even if you want to speak a little bit more
22:27
about L three or even L four, how that might impact them. Well,
22:33
the first implication, hopefully is that we will improve the experience of employees
22:38
around the world and we will help it you will do a better job also
22:44
and improve the expence of IT teams. And my vision and maybe my hope
22:52
I could call it, is that we are not going to replace at one,
22:56
two or three teams, but we are going to change the way they
22:59
works, that what they do can be more impactful and more meaningful because it's
23:04
the same with every one and too, you know, like next thing close
23:07
to automate a lot and reduce the meantime to repair. So in the end
23:14
there's probably a lot of the work that today every one and the two teams
23:18
are doing, which you know, thanks to nexting can disappear and same FOLK
23:26
three with what we've just discussed. But once again, I think what it
23:29
will allow those people is not to replace them, is to make their job
23:33
difference and more meaningful. And hear what I mean is I think that those
23:38
people could have a role which becomes more of an advisory role than just you
23:44
know, fixing stuff. And why I'm saying this is because you know,
23:48
with next thing and the data that we provide, and you know, as
23:51
we said, we're going more in mind, this direction will provide more and
23:53
more of this data. Imagine that those people, they will have access to
24:00
all these cross customer data about you know, they will be able to say
24:03
which devices, which device models work the best in which circumstances, which applications
24:08
are the best for this or that use case. You know, which headsets
24:12
are the best when you have a zoom call versus a team's call. And
24:17
imagine how powerful this is because those people in the end, even if they
24:21
have power incidents to fix, they will be able to really fine tune improve
24:26
the experience by becoming advisors and fine tuning and saying, you know, you
24:30
should you you're a sales rep, you're using this type of applications, You're
24:37
having calls all day long, you should have this type of device and this
24:41
headset and use this application you know on this Chrome version or Edge or whatever,
24:48
And I think this is really powerful. All these also by keeping into
24:52
accuntoc cost and you know it's it's really it comes to this use case of
24:57
personas that everyone and is dreaming about, which everyone is saying they do,
25:03
but in fact nobody really does it, so personalized deity and being able to
25:07
say, for this person, for this youth case, this is the best
25:11
device, This is the best equipment, the best headset, and the best
25:15
application, the best configuration that you can have to do your job. So
25:19
I hope that's really well where we're going in the future. Absolutely fantastic descriptions
25:27
and thanks to Oriana for handing me on a plate. What I think is
25:30
a perfect title of this episode, the Encyclopedia of Decks, I think,
25:36
and very very well earned. Indeed, before before we say goodbye, we
25:41
are asking all of our guests at the moment to talk a little bit about
25:45
what they what they enjoy most about their role as Next Thing, because what
25:49
they what they appreciate most about the culture of Next Thing and what they do
25:55
and very interested in your case to hear what you've got to say about it.
26:00
Yeah, well there's a lot to say once again, but I think
26:03
for me, one of the most of the best thing working at texting is
26:07
working with talented people. So that's really cool, you know. You every
26:11
day I wake up and I know I'm going to interact with, you know,
26:15
people who have high skills also who are fun and it's super great to
26:19
work with people like this and who are gonna, you know, teach me
26:22
things. But I also like making an impact. We really see how we
26:26
help it teams, and you know, we improve the digital experience of millions
26:30
of employees around the world. And because we really get things done, you
26:36
know, we really is every five weeks. We really see all these new
26:38
features, all these new things that we allot to do, and we get
26:41
this feedback. We have this fit backgroup working very well. We see how
26:45
this works and how how it's helping, and I think we are not afraid
26:51
of challenges. That next thing, I think everyone would agree. So a
26:53
few years ago we said, you know, we are going to redo the
26:56
whole products from scratch and move to a SaaS product and you know, from
27:00
unprem to SAS and we've done it. Now it's you know, it's a
27:03
it's a done deal and the new product is it is so great. So
27:06
it's really cool to work in this type of environment where we're not afraid of
27:10
you know, doing big bags, were working on big things and make and
27:14
making them happen. And there's a lot of challenges, which is also a
27:18
cool part of the job, because otherwise it would be boring. And if
27:22
I talk about my role specifically, I like also interacting about you know,
27:26
interacting with various stakeholders, so from the most technical people to customers and you
27:33
know, legal sales and so owners of us and and that's also very cool.
27:37
And the last thing, which of course is super important, is to
27:41
have fun. And I think that next thing we have fun, people are
27:42
nice, we know also how to celebrate and yeah, super important, very
27:49
very well. Pokam, thank you so much for coming of a tech show,
27:52
coming on reality. But I should say it's been a pleasure. Thanks
27:56
for having me. Thanks everyone. To make sure that you never miss an
28:00
episode. Subscribe to the show in Apple Podcasts, Spotify, or your favorite
28:04
podcast player. And if you're listening on Apple podcasts, make sure to leave
28:08
a rating of the show. Just tap the number of stars you think the
28:11
podcast deserves. If you'd like to learn more about how next Think can help
28:15
you improve your digital employee experience, head over to nextthink dot com. Thank
28:19
you so much for listening until next time.