Atlassian IT Transcript

gistfile1.txt

0:03
[Music]
0:20
hi hello everyone we'll sit down in a second i'm kind of impressed you voluntarily
0:26
came to a topic about it and i think the farthest room you had
0:32
to walk to um so i'm assuming that if you're here you really want to be here i'm jenna klein i'm going to moderate a
0:38
session i'm going to introduce my team and we're going to go through a few questions we have but we want this
0:43
to be interactive we're going to have a few questions for you all as well all right thank you for joining
0:48
okay first this is a session about managing a
0:54
modern i.t organization and the way that we run atlassianit we like to think of as modern because we
1:00
have a few different philosophies that we bake into our decisions and we've talked to a number of
1:06
customers over the last year about this concept of modern i.t and there's three that really resonate with a lot of
1:12
customers and we're going to get into those today but first let me introduce the panel because i do have experts on here that
1:17
are going to help us navigate some of these big topics okay first we have mo
1:23
who is our head of intelligent automations which includes automation as well as integrations and
1:30
he runs a it innovation team
1:35
and automation for us as it is probably for a lot of you as well has been a
1:41
important investment for us last year we hit our big hairy audacious goal of
1:46
automating out a hundred thousand hours of work in atlassian with moe's team
1:52
and we're going to talk about that a little bit more in some future um but most of the the work that we did was
1:58
based on low code and no code platforms so that's mo next up is sunny he's our head of i.t
2:05
strategy and business operations he looks after our investments our hiring
2:11
plans and a lot of his current work is really around managing our ever-growing
2:16
sas application we're mostly a cloud-based i.t team and the management of that falls under uh
2:23
sunny and his super capable team and then at the end here we have phil
2:29
and for anyone who's worked in it for more than a few days you know that just deploying technology does not make it
2:35
successful you need someone you need to make sure that it's the tech it's the people it's the process phil was our
2:41
first business architect that we hired at atlassianit to make sure we do just that he sits in between the tech and the
2:48
business helping translate business goals into it investments right there we go well let's get into it i
2:57
am going to bring up sorry i think this is hidden there it is the three three topics we thought we'd
3:03
get into today we have a lot of choice there's a lot of different things we could talk about about managing a modern
3:08
i.t organization the three that we chose today are around managing i.t versus shadow i.t
3:17
second is running i.t like a product organization and then the last is baking innovation
3:22
into our dna before i get into this i'm going to ask
3:27
quickly in the room by show of hands um how many of you are actually in nit
3:34
okay i wasn't expecting that so almost everybody is here is nit that's actually great we're going to have some nerdy i.t
3:40
jokes so i just wanted to make sure you all got them um i also because i t can really vary
3:45
company to company and the profile of the company can impact what it looks like i thought it'd be
3:52
good if we just went over a few fun facts about rit organization the first as i mentioned cloud native we
4:00
mostly buy cloud we don't have a cloud migration strategy for uh it because we
4:07
are cloud we've been that way for a number of years we have very few on-prem things so
4:12
we're managing a cloud portfolio and in that portfolio we have about 600 apps
4:18
for the whole company and we actively manage around 270 of those
4:23
so we do have a bit of shadow i.t and we're going to get into our thoughts on shadow i.t
4:29
we are global company and we're a global i.t team so the way we manage i.t is
4:34
around hubs or geographic hubs we have a distributed workforce as a company we
4:41
hire anybody anywhere but we consider it as having three dominant hubs where we like to co-locate
4:48
when we need to or make sure we have staff and that's in the u.s india and
4:54
australia we are a product-based operating model that means that everything that we deliver from it is
5:01
thought of as a product that we build or purchase or integrate or make available
5:07
to our business our corporate teams as needed it's a little different than i think some other uh operating models i t
5:15
that i've worked with in the past like service management models the product operating model is a transformation
5:20
we've been on and we're going to touch on that in a bit we are a best of breed i.t so we buy best of breed we don't
5:27
single source on a platform for example we want to make sure that we have what works for our business and we prioritize
5:34
the experience and the collaboration of our teams with the tooling that we purchase and then last i think it's obvious but
5:41
we have the awesome opportunity of being an i.t team in a software company that is growing
5:47
is global that is remote and has an awesome culture and set of values that we really love so that's
5:54
where we're coming from let's get into our questions let's talk about modernity guys all right yeah do
6:00
it smile it's a good topic come on the first one i got a poll for the audience another one for you
6:06
and this one is going to be around shadow i.t so in general how many of you think shadow i.t
6:13
is is evil you really try to clamp it down you're trying to get rid of shadow i.t it's starting to overwhelm your
6:19
company shadow i.t on the more negative side okay and then how many you think shadow it's
6:25
awesome three and one works for me okay so
6:34
um we actually do have a little bit of a philosophy on shadow i.t i we've touched on it on a couple of topics if you went
6:40
to the panel that i was on with my colleague molly hellermann that was moderated by a forester
6:47
uh analyst chris condo we talked a little bit about shadow i.t we do embrace shadow i.t but
6:53
that doesn't mean it's it's the wild wild west of shadow i.t so let's talk about what we actually think about it
6:59
and i'm going to start with phil how do we define shadow i.t
7:06
sure so you know our definition really isn't all that uncommon it's i.t assets that aren't owned by it right so real
7:12
big brain stuff here but you know when we look further into that we we kind of have a pretty simple
7:17
classification mechanism we use so to get the easy stuff out of the way first there's a big chunk of tech in atlassian it's owned by our product engineering
7:24
teams now these are technical competent folks that are very very clear in terms
7:29
of how they want to manage those assets we sort of let that go and let them do their thing with those that said attack
7:35
the other bucket a little bit a little bit harder to nail down is those apps that are owned by the business and you may be thinking like why would you want
7:41
to let the business run a piece of technology you know there's a lot of these niche sas applications that are
7:46
coming up and when we see you know an app that doesn't necessarily need to integrate with a whole bunch of stuff say you
7:52
don't need sso integration fruit for some reason or that the business team has you know
7:57
center of gravity of skills around that app they've got competent folks who know how to configure them we'll generally let them do that within a set of guard
8:03
rails there's a sort of a third class of shadow i.t this sort of new realm of shadows that's popping up that we're
8:09
starting to get a handle on looking at stuff like app plugins browser extensions things like that a
8:15
little bit harder to nail down we've got a few different mechanisms that we use to keep an eye on those but really our philosophy center is around visibility
8:22
making sure that we've got an awareness of all of those things that are out there there's a couple different ways that we do that first and foremost we partner
8:29
with a number of different teams across atlassian working with procurement i.t security risk and compliance and privacy
8:36
to make sure that anything new that comes through the door at least gets a look from all those teams right we don't want to invite new tech in that's not
8:42
secure you know some new tech that's not complying with gdpr so we make sure that those kind of non-negotiable those boxes
8:48
are checked before we let it through the doors in addition you know that affords i.t a great opportunity to make sure that you
8:54
know beyond just sort of those non-functional requirements we get a good look at making sure that functionally from a sort of a business
9:00
need a business value perspective that we're hitting the mark so if we see somebody that comes through with a tool that maybe we already have a
9:06
few of we can start to direct them toward those tools to avoid further proliferating the ecosystem
9:12
one other thing to call out if you missed the talk this morning my colleagues over here pete and leah did a great talk about how we employ sas
9:18
management as another discovery mechanism to make sure that whatever we already have we can sort of pick out and
9:25
find those apps and make sure we got a good handle on that so you know scanning our financial records making sure anything that folks are paying for we've
9:31
got a record for we kind of sniff that out we can make sure that you know it's compliance secure and meeting the needs
9:36
of our business yeah so if bring it out of the shadow into the light
9:41
i used to have a boss who would say sunlight's the best disinfectant which is kind of harsh uh all right so
9:49
um sunny i think one of the reasons a lot of it teams and
9:55
i know all of us have been on itunes in the past where shadow i.t is really something to contain because it can be
10:01
expensive right aren't we managing or aren't we worried about the cost how do we manage that
10:06
yeah you know from us we're fortunate from a strategy perspective we really look at this from a bit of a different
10:11
mindset while we do absolutely observe and manage the spend across our portfolio of sas applications cost isn't
10:18
necessarily the leading decision driver when it comes to purchasing sas applications we lead with enabling the
10:23
business with the needed capabilities while ensuring all security requirements are wet we do have a governance
10:29
framework in place to help ensure that our applications are indeed providing the expected business value from both a
10:35
capability and leverage perspective if we find that the applications aren't meeting the business value expectations
10:41
then we will evaluate options to potentially rationalize or even retire the applications um to share a bit of an
10:47
example when we first went into a fully remote work environment as a result of the kobit shutdown we wound up
10:53
implementing multiple solutions to enable virtual whiteboarding capabilities now all of the solutions
10:59
they provided similar capabilities but we just had teams across the company that preferred one over the other and
11:04
you know during that somewhat difficult time what we did not want to do was negatively impact experience or
11:09
productivity for our workforce so we went ahead and implemented all the solutions and as part of the upcoming
11:14
renewal we will evaluate and assess whether they're all providing the expected business value and if we find
11:20
that one or more are not then we'll assess and potentially retire some of those applications so to bring it home
11:26
our strategy is to focus on business value versus cost when it comes to managing our sas portfolio
11:34
um so mo it's one thing to say we have visibility and we are managing costs and
11:40
other things but is there actual goodness in shadow i.t we should definitely embrace shadow i.t
11:46
there's too many requests coming into it from the business to be able to service everything
11:52
industry analysts are telling us by 2025 70 percent of all the technology created in the enterprise is going to be
11:58
done by non-engineers and maybe someday you know they will have ai writing code but in the next
12:06
short term next few years it's going to be business technologists that
12:11
are not in it creating that technology so as i.t we need to enable that technology creation but also
12:18
provide security and compliance to minimize risk
12:24
who used to be a jira admin from like 10 years ago 10 plus years ago okay so you
12:29
might know what i'm talking about so back then you know to get jira set up you had to download a zip file unzip it
12:36
edit a bunch of config files get the database running get the web server running and then you could start using
12:42
jira we used to send people a t-shirt if they managed to get
12:48
through that right of passage the dragon slayer t-shirt if anyone remembers that from back then
12:54
um fast forward to today to get you spend more time using jira than
12:59
setting it up because you just fire up a web browser and you're ready to go so we want to take that analogy and apply it
13:05
to all of the internal tooling that we create and a recent example i can cite is based on the foundation org you may
13:13
have seen them in the expo they have a booth where you can make solar panels for
13:18
developing countries if you haven't seen it you can spend 30 minutes volunteering with them they coordinate a lot of our
13:24
volunteering and charitable efforts like donations for the for the company and they don't have
13:30
any engineers they don't have any part of it that's dedicated to support them
13:35
and they had this program called engage for good where they wanted non-profits
13:40
to offer volunteering opportunities for atlassians that had particular skill sets so coding design project management
13:48
etc but we're talking about hundreds of non-profits
13:53
matching with thousands of employees that are last seen globally so manually if manually doing that would have been
13:59
thousands of hours of effort but they managed to automate all of that by spinning up their own jira service
14:05
management instance creating the form for the non-profits to fill out and offer their opportunities and using some
14:11
no code integration to match that with alaskan skill sets and then notified them on the back end and they did this
14:17
with minimal help from it and if you want to read more about that particular example and others like that
14:24
that it's helped enable shadow i.t you can go to the connected cio microsite i
14:30
think there'll be a qr code later to visit that site so shadow it we want the light we want
14:37
visibility so it's get it out of the shadows make it visible make sure we're managing the cost as
14:44
needed but don't sacrifice the experience or or the value and then um
14:49
there can actually be goodness in some shadow i.t and enabling our business to do what they need to do i mean i i
14:55
think i mentioned if you went to the talk this morning with molly that one of the things that i've seen is that the
15:00
a lot of innovation happens outside um traditional tech or big platform
15:06
plays right some business teams have a very niche problem that need to be solved and they can find an app to do
15:11
that and so if we allow them with the guard rails that we need to bring that in quickly
15:17
to solve their problem if it gets to be a prolific use or something that we can structure with a better agreement we'll
15:23
do that but more importantly our business teams are really able to pick up some of those small uh tasks that they would are
15:30
waiting in line for it to solve um and they can do that quickly and at atlassian we are a software company and
15:35
we we have a pretty tech savvy workforce but a lot of these low code no code options that are becoming available are a great
15:43
opportunity to take on some shadow i.t without you know scaring people
15:48
at the company with by embracing it fully you don't need to tell them that atlassian told tell them to love
15:56
okay shadow i.t let's move on to no that one let's move on to
16:01
all right another poll how many of you have in your it teams a role called product manager
16:08
or product owner okay couple yeah interesting
16:14
um we have been on a journey for the last couple years well about 18 months i guess in atlassian i.t to move to a true
16:21
product-based operating model we've identified 50 products i think we have right now in
16:27
our portfolio we we group those and we manage those at 11 group levels and we've built a
16:35
a core team of roles for each product so we know what products we have what roles
16:40
we need in those products who staffs those um and we're moving into how to budget and invest those how to measure
16:47
that etc it's been a it's been a journey we were a bit more of a mixed bag of
16:52
service management and some other ad hoc styles we were always deploying agilely but we wanted to make sure that we were
16:58
focusing just like with our shadow i.t philosophy on making sure that we're delivering an awesome experience and
17:03
putting people in the role of a product manager helps build that mindset that that's what we're trying to achieve
17:09
so a question for sunny um
17:15
sunny sonny and i have actually been on a team we were in more of an itsm model very we were very strict itsm model uh
17:23
in a previous life from your role in bizops or business operations right now
17:28
since you've seen different models of running i.t what can you tell us it's different about running a product-based
17:33
operating model yeah sure so the differences are subtle but very important for us
17:39
as jenna mentioned we started the journey of implementing the product-based it operating model a little over a year ago
17:44
and one of the first decisions the i.t leadership team had to make was whether we were going to go forward with a
17:50
product a service-based model now both models do carry forward similarities such as organizing all aspects of i.t
17:57
around a common taxonomy and framework but given that atlassian is a craft-based company focused on building
18:03
amazing products we wanted to make sure as an i.t organization that we led with the product mindset when working with
18:10
and enabling teams around the company so for us the product mindset truly is the key differentiator between the service
18:16
and product models yeah yeah that mindset shift has really been
18:22
um a big one for us um mo when we were first kicking this off i
18:28
remember we were laying out the philosophy or guiding principles that we had for product operating model and
18:33
one of the ones that it leadership had the strongest opinion about is they absolutely wanted to build out
18:39
autonomous teams as me as a someone who's been in i.t for
18:45
a while autonomous can sound scary uh to say these teams can go and operate
18:50
how they want to operate what's atlassianit's perspective on building autonomous teams yeah so the
18:56
autonomous teams they can make decisions very quickly adapt to change and hence
19:02
deliver value very quickly but the product managers act like the
19:07
glue between all these different autonomous teams so each of the it teams that own a product will have a product
19:14
manager paired with the engineering manager and sometimes also try out with a designer if there's a user experience
19:20
component if i take the example of the automation platform like we provide a platform for
19:27
all teams to automate their own work but there's still a need for a product manager to work out how teams want to
19:33
use it what features they want to see start filling up that jira backload for the engineering manager to burn down
19:39
but that doesn't mean the product manager decides everything that that engineering managers and the team is
19:45
going to do because the engineering team still needs to worry about a lot of the non-functional requirements so analytics monitoring
19:52
alerting all the ups and incident management and together that sort of that partnership
19:58
ensures that the automation platform is not only usable and provides a good user experience but
20:04
is also compliant secure scalable performance and reliable
20:11
uh which is all good but it does sound expensive sonny it's a mo is expensive in general but
20:17
like this philosophy also sounds expensive how how are we um thinking about managing our budgets and our
20:23
investments with this product operating model yeah i mean it's absolutely going to require a mindset shift across the
20:29
spectrum as part of our journey of implementing the product-based itr operating model one of the first things we're going to
20:35
do is begin tracking our financial insights based on the product taxonomy and we will also begin holding product
20:41
managers accountable for the financial performance of their products as a part of that we will also begin prioritizing
20:47
investments based on products versus projects and within each product we're going to categorize the spend into four
20:54
categories the first category is incubate which are investments towards innovation type of initiatives the
21:00
second category is transform which are investments towards the initial build in implementation of capabilities and
21:06
platforms usually within a single product area the third category scale which are investments towards scaling
21:13
those capabilities and platforms across multiple product areas so that we can leverage common capabilities across the
21:19
entire stack and the fourth category is bau or business as usual which includes
21:25
ktlo keep the lights on type of work which also includes addressing technical debt as well as small features and
21:31
enhancements type of work that aren't quite large enough to be grouped into initiatives or projects and you know i
21:36
mean we're pretty confident or at least i'm pretty confident that this framework will help ensure both accountability and
21:42
transparency of all of our investments across our portfolio products yeah and each year with those four
21:49
categories we we set some goals for our product managers to look at their
21:54
investment so a percentage of their portfolio should be going towards each and we try to keep
22:00
that business as usual that ktlo work lower that gives the product managers um
22:05
with that goal in mind they can then make more investments in what they need to do to introduce new capabilities for
22:11
their product and we'll talk about that as an it leadership team each year about what do we think the right mixes do we need to
22:17
push more on one thing or another that of course would change with business models acquisitions that we make investments
22:23
etc but the framework is simple enough where it helps team understand how to think about where they should put their
22:29
investments um and then we can roll that up across all our products at the i.t leadership level and be able to see how
22:36
we're making investments each quarter um so these investments that that were
22:41
um talking about in these autonomous teams were running phil that's on the i.t side
22:47
what does this mean for how you and our other teams are working with our business does this change the way we
22:52
engage with our business teams you know it it's changed a little bit but really in a positive way you know it's kind of
22:57
simplified our engagement model right we've got a consistent set of roles organized in a very consistent way
23:03
across product groups and individual products so it makes it much easier for our business ops teams our business
23:08
partners and more day-to-day engagement to know exactly who's accountable for what right they've got their product manager they've got that technical lead
23:15
so it's sort of streamlined that quite a bit especially if you know you come from more of a mixed model as jenna had mentioned where you know certain groups
23:21
organized in certain ways that consistency that simplicity has paid dividends already now when you're talking about more
23:27
transactional engagement right somebody wants to log a feature request or hey my laptop broke
23:32
one of the other things that we did was unify a whole host of different jira service desks that we'd been accumulating over the years and big
23:38
shout out to our workplace technology team this was was a bit of a bit of an effort
23:44
and what we found is by having that single front door for it it gives us a
23:49
totally clear pane of glass into which business customers are coming to us for what you know we can report very very
23:54
simply given that we've got that common instrumentation across all of our it teams now it's been in place for what about a year or so almost a year yeah
24:01
yeah we've started to see quarter over quarter improvements in terms of you know resolution times coming down you know improved csat from business
24:07
customers who are getting that support getting answers to those questions much more quickly so we've seen a definite move of the
24:13
needle there and if i could be selfish for a minute just talking as a business architect where you know we generally get plugged
24:19
into you know larger cross-functional problems that are going to require a whole host of different product teams by having that kind of predictability
24:26
across our it model it affords us a strong accelerant in terms of assembling the right groups of products to to
24:33
tackle whatever that business challenge might be so so definitely a big fan here yeah and the service desk one is
24:39
interesting when we were starting down this journey of product operating model i just wanted to look at basic ticket data um and health and things across i.t
24:46
about two years ago and that's when i learned we had 17 service desks in i.t and we put the we put the burden on the
24:52
end user to go figure out which service desk they were supposed to go to to get their ticket help and then each service desk was using
24:58
their own issue types and we had 42 different issue types in i.t and we thought well we probably should clean that up right so we got 17 actually 14
25:05
and we eliminated few we got those 14 down to one and then um as part of that that service
25:10
desk is a product itself like that they should be thinking of these things how do we make that experience better for
25:15
the company not just um you know for a few people who might be managing tickets for example and uh one of the things we
25:23
did is not only combine the service desks but we gave a single entry for everyone in atlassian to go get help
25:29
from it like we'll take the burden of knowing of of routing those and where they need to go so they have one place
25:34
to go we have one portal that we're managing and then product managers get all the
25:39
information that they need on on what requests and other things will be coming in that's that's been a big big change
25:44
for it yeah that was a journey lots of happy business users lots of happiness users
25:51
lots of questions for my t people okay um let's let's move into the third
25:56
tenet here and that is innovation that's a hot topic um
26:04
atlassian of course loves innovation um it's something that we fully embrace uh
26:09
and even in i.t so but that's not always the case and i'd like to get a spectrum here of
26:15
if you believe your i.t team at your company is paving the way for innovation
26:20
at your company okay some tents demands like okay
26:26
you thought of a few examples in your head they were like i was kind of innovative yeah um
26:32
this is a tough one i mean our people want to work on innovation right the innovation is the interesting work they want to innovate they generally work in
26:38
tech because they really like tech so they want to be in that space but i.t is busy i get it like there's
26:44
more work than we have time or capacity for it often takes an investment
26:50
so we've we've tried to navigate that ourselves at atlassian i.t to make sure that we're we're providing um
26:57
innovation opportunities and really trying to bake that into our dna at it so let me i'll start with mo because you
27:05
actually lead an it innovation team and i'm guessing this is where all the
27:10
innovation happens right right in your little team so i think we can all agree that innovation's a good thing right we're
27:15
coming up with new ideas inventing new ways of working employees are more satisfied because
27:20
they get to work on things that you know they thought of but the problem we all have is finding the time to do
27:27
that innovation and at atlassian there's a number of innovation programs the most popular
27:33
being ship it which is a quarterly hackathon and that gives you 24 hours
27:39
as a whole company to work with whoever else you want on whatever you want
27:44
and that new products ideas have come out of those hackathons but it's only four times a year once a
27:50
quarter and other teams run innovation weeks which is a bit more you know gives a bit
27:56
more time to make something a meaty innovation at the end of it but again they usually run four times a year
28:04
but what we found in it the best way to get more innovation going is build it into your it your team's operating
28:11
rhythm so just do innovation all the time and how do we
28:16
do that um i mean first we have to encourage the teams to keep across all of the trends
28:22
the hype cycles the disruptive tech for their particular domain and
28:28
my team tracks that on confluence using a tech radar type format on a template
28:33
and as you're watching those trends some ideas will come out on potential use cases within the company we capture
28:40
those ideas in regular ideation rituals biweekly and capture them on trello
28:47
when it gets closer to prioritizing those ideas and you actually want to start evaluating technology or doing a
28:53
proof of concept that's when we start thinking about how do we get the capacity the spring capacity to actually
28:59
do that work and you know sunny was talking before about how we carve up all the different uh
29:05
buckets of budget so if your current sprint capacity is 80 new project work and 20
29:12
business as usual you know try doing 70 percent project work and carve out that 10 percent for innovation
29:20
and put into your sprint backlog like any other work but these innovation projects the best ones
29:26
show value quickly or they fail quickly and failing is not a
29:32
bad outcome for innovation project um because there'll be some learnings there like technology was not ready
29:38
maybe the company wasn't ready so it's good to do a retrospective harvest all of those learnings
29:44
um what went well what didn't go well and park it off to the side and maybe
29:49
another team in the future will come and take it you know that next next iterative step and get it closer to reality and we used
29:57
to have like jenna mentioned i run an incubation team innovation team so that team used to spend all of their budget
30:03
on incubation but over the years they started to work on bigger projects so they're doing a bit more transformational work now and all
30:10
the other teams are now spending more of their budget on incubation yeah yeah and the failure i think is an
30:15
important point like it's it's it's not necessarily failure in that people failed it's failure that okay maybe this
30:21
innovation wasn't exactly what we needed to be but the the learning is super important i think bob and scott talked
30:27
about this at the keynote this morning it's we don't want to rake anybody over the coals for failing it was you tried and
30:33
we learned and now let's try something again is a great way to look at it however it
30:38
it can smell expensive sonny so again what are we doing at the
30:44
investment level in our portfolio to balance our innovation costs yeah and i mean when you hear about the
30:50
breadth and scale of innovation as mo described it it can feel and sound a bit expensive um for us we take a blended
30:56
approach when it comes to investing in innovation as i previously mentioned we categorized as spend into four
31:01
categories incubate transform scale and bau and mo alluded to this too a bit as
31:06
well but one of the levers we use to invest in innovation is by reducing the
31:12
amount of our investment in spend in bau type of activities now you might be wondering well how can it sounds easy to
31:18
do that but how can you actually do that um you know two ways that we've been able to more consistently do that within
31:25
it is one through the automation of work and two is by
31:30
scaling single capabilities or platforms across multiple product areas so that we can
31:35
leverage those shared capabilities across um outside of that you know innovation is a big focus for us so we will
31:41
prioritize net new investment as needed to fund innovation i mean you know i think we have a really good example of
31:47
this moe's automation program is mo said it started out as a very small incubation program and now it's
31:54
gone all the way through that model and now we're kind of at scale even starting to carve out some bau in some areas um
31:59
any learnings or inside smell you'd like to share about that um i mean one thing like from the outside in it looks like
32:06
that automation platform popped up overnight but like when i go back and look at all the
32:11
different tech evals and pocs we ran there was probably 30 different iterations of failing before we got to
32:16
the right right mix of different products and solutions we needed to provide automation across the whole
32:22
company yeah i remember when we had the conversation at um our it leadership it was hey
32:30
we should probably look at rpa okay do we have any money yeah we'll put some money and who should we give it to we'll give it to mo
32:36
and that was uh that was our investment that year for uh innovation it was for rpa at the time
32:42
and it took off it really did but we had to make that investment and make sure we told somebody your capacity is
32:48
going towards experimenting with this with this tech uh okay so we do a lot of innovation in
32:55
i.t obviously phil but we work with the business too how are we making sure that we expand
33:03
innovation things that we're looking at anything that it thinks might be promising for the business how are we
33:08
expanding it across the company you know absolutely and it's both across our business partners as well as with our
33:14
product organization as well so starting with the business partners you know we've got as mo mentioned sort of
33:19
this fail fast mentality uh one of the new things that we've spun up in our atlassian on atlassian team or customer
33:25
one is something we affectionately call simcity which is a simulation of rit environments so
33:31
think of it as you know you've got yeah we'll just say a playground to fail when
33:36
you're talking about exploring new technologies you know exploring new integrations connecting different technologies so just to give you an
33:43
example you know we use workday as our hris you know it's a small platform some people some people may have heard of it
33:49
and we'll just say our talent acquisition team is looking to upgrade you know how we're doing our hiring they they want to explore a couple different
33:55
hiring tools and really quickly understand those that actually are up to snuff and and actually meet that sales
34:00
pitch you know doesn't integrate well is the data you know exchanged in the appropriate ways you know is a
34:05
delightful user experience so rather than going through you know a super lengthy rfp process and you know finally
34:11
getting the right tool in the house and then plugging it in and then figuring out you've got a couple things that aren't quite right
34:16
we can quickly spike on multiple tools and sort of get to that point of okay it is working or it's not
34:22
and then sort of make that larger you know financial and contracting decision and bring in the right tool and feel better about that but it exposes those
34:28
failures in a really really fast way so if you've heard of the concept of a digital twin it's kind of similar to
34:34
that so a very very cool thing that we're working on there in terms of you know supporting our business partners as they
34:39
continue to have this insatiable appetite for new tools and technologies uh and in terms of how we're working
34:45
with our product teams that same team inside atlassian has a second offering called our mock rfp
34:51
so we're big fans of our own tools shocker so often we'll you know do our best to
34:57
employ those tools where the use cases make sense and one of the services that that teams offers is actually providing a very
35:04
discreet output to our product teams to say hey you know we're looking at using say jira service management for this use
35:09
case we're gonna actually stack that up against our competitors and pretend like we're running an rfp against it
35:15
sometimes the outcome is pretty positive right you say okay cool our tool is absolutely going to work for that uh other times it is not
35:22
but when it's not we've got that product feedback loop to sort of share all those results with our product leadership our
35:27
product teams so they can start to close those gaps and make our products even more competitive in the market yeah
35:32
yeah we had to share we did it started with the premise this this company on company program it started
35:38
with the premise of if we weren't at lassie and i.t would we still choose atlassian products
35:45
and we developed a number of tools we wanted to look at a tool just agnostically like we would look at any
35:50
other vendor when we were making an assessment and we have a couple of tools the ways we do that the mock rfp is one of those
35:57
we use our our rfp our request for a proposal framework that we use for other vendors we use that against our own
36:02
tools sometimes we look at our tools and we decide we wouldn't actually use them
36:07
and then you have to go in front of the executive team and explain that but they you don't sleep the night before but
36:12
it's actually great data right we we're operating under the premise that if it doesn't work for an enterprise it like
36:18
us it's probably not working for enterprise i.t like you all and we want to make sure that our
36:23
products allow everyone to have an it team that runs awesome and so we're
36:29
making sure that we do that with our own products sometimes in use cases that
36:35
the product wasn't built for right and getting that feedback and then getting that feedback back to the product teams
36:41
in a structured way that they understand which is another thing that really helps the product our product operating model
36:47
allows us to have the same language as the product teams so when we're working with them we're talking to them and
36:53
using the tools and techniques that they use to build products for the company that we're using to
36:58
build inside corporate atlassian but we can get our requirements back into them quickly
37:04
so it's been a great opportunity for us to innovate in i.t and atlassians are passionate about their products and they
37:10
want us to be able to expand and so we can take those use cases to our teams across the company as well
37:16
okay awesome i just noticed the time and we want to leave some time for questions right we hit all of ours all
37:22
right if anybody has any questions for anybody here mostly these guys
37:29
i'm happy to have them answer you okay we have one here yeah we're getting a mic to you
37:35
thank you um i just have a question about uh product managers is uh could a single product manager in your guys's
37:41
environment manage multiple products or are they just assigned to a single product that they're managing
37:47
it's so i think the question you're asking is do we is it one product manager to one product exactly or is it
37:52
one product manager to multiple products exactly it can it depends on the product so we have some uh very complex
37:59
platforms that we're running that's going to take uh layers of product managers so we probably have a group product manager overseeing that senior
38:06
product managers that are overseeing a significant portion of that portfolio and then product managers that might run
38:11
a feature set um it's smaller whereas other products might be a lot simpler the portfolio might be simpler and they
38:17
can especially if it's just sas and they're mostly navigating with companies that they would have one product manager
38:22
to one product thank you
38:32
so in your product-based model how do you handle those services or products that are purely
38:39
service such as i.t support or hardware yeah that it's a great question
38:45
in fact this was a heated debate when we were going down this path in it
38:51
leadership it's amazing the stuff you can like spend hours talking about is what's a service versus a product
38:56
uh we even toyed with this really confusing concept of uh product services product of the services
39:03
product as a service um that really that failed to talk about failing fast that one didn't go very far
39:08
thank god so we do we did in the end decide um that we we do have a few what
39:13
we call shared services um but in general it's it's a few it's
39:18
an exception to the rule our business operations team would be one example sonny's team does provide a service um
39:25
on how he manages things they have a tool and portfolios to do that but it is a service that's provided to everyone
39:30
but mostly we are products even in the it
39:36
support that is a product right like someone's looking should we go from 14 service desks down
39:41
to one should we also build a single point of entry how is the knowledge base tying into these and we use confluence
39:46
for our knowledge base so who's thinking of the experience of the knowledge base and working with the confluence team on a roadmap where we might want to take
39:53
advantage of new features that are coming or influence the product through our company on company program it's really once you say
39:59
your service is a product you own a product a service is waiting
40:05
for someone to come and ask for something or give direction a product is someone who's motivated to get more
40:10
customers using their product and so it really helps we want to be more product based than where we are but we're not
40:16
going to force fit everything if it doesn't make total sense
40:27
hi um i was one of those few people that raised their hand when you asked her do you loathe it or love
40:34
it i said i looked oh nice shadow id that is so um a two-part question
40:40
um i raised my hand because the fact that shadow i.t means there's
40:45
not enough oversight or visibility from the i.t side so part one of my question is um are there
40:52
any recommendations as to as ita walls and more and more
40:58
systems are decentralized and people start managing things on their own is there any run book that you guys are
41:03
using that has helped you to ensure there is still that accountability and ownership with respect to security data
41:10
protection there are so many other areas that uh centralized iit and security teams do a little bit better in my
41:16
opinion today at least in in companies like splunk i'm from splunk by the way so um versus how
41:23
others do it that have been successful with it i hope this question is clear yeah so i think your question is i mean tools
41:29
there's proliferation of tools so how are we making sure that the tools that we do have visibility into those tools
41:35
are we using platforms are we using any detection and how are we managing those yeah more on the governance side as well
41:42
basically so if you don't run those systems at the same level of scrutiny that ite would usually do a matured id
41:48
organization would do how do you ensure that decentralized teams can still manage their systems with
41:54
enough oversight and governance is there a handbook or has a has there been some
41:59
good best practices that i've worked for at last seen in their data there has been some good but i'll answer a few and
42:05
then i think these guys probably have an opinion on this too also i will say that um
42:10
earlier today my team had a whole topic on this called peeling the sass onion that will be available as a recording if
42:16
you want to hear more they're the experts on it okay there's a lot more into this topic but it's a great question in that the tooling keeps come
42:23
up how do we get visibility how do we govern um things so we have a philosophy that if
42:29
you need it go buy it and but once you buy it doesn't mean we don't want visibility into it and so
42:35
when the purchases are made we do have a very lightweight governance that looks at the purchase and makes sure we have
42:41
some basic things in place and then once it's in-house we're mapping that application that we purchase to our
42:48
capability model and that's the product managers then can have visibility into everything that maps up to them
42:54
the tooling behind this and how we're trying to automate that so it's not a big owner it's governance process with a
42:59
lot of spreadsheets it's something that we're maturing right now and we have a number of tools in place to allow us to
43:04
do this but as you can imagine i mean 600 apps all the data fields and all the
43:10
information that can quickly get out of date so our goal is to using our own tools eventually get this automated so
43:15
it goes from purchase to understanding that it aligns to this portfolio for the product manager
43:22
i will add anecdotally again back to the product operating model my question as an i.t leader if i'm seeing
43:29
some it product manager's portfolio growing with apps that the business is purchasing is why
43:36
aren't you need to have a closer relationship with your business that you identify the gaps before they identify the gaps and
43:43
go buy something because that's the real problem that's going on there the it team isn't in front of the business
43:50
problem and helping work and identify the value before the business just gets frustrated and goes off and buys their
43:55
own thing anyway that's the culture side of it did anyone have anything else yeah i mean i'll just say there's not one solution or process that
44:01
fixes this entire problem it's an ecosystem of solutions and an ecosystem of processes that we employ and as jenna
44:07
mentioned we're still maturing our own practices around this yeah and
44:13
uh one specific area where we're actually putting a little bit more accountability on our it teams is through our solution architects
44:19
so i don't want you guys to think that just because it doesn't own it we don't care the solution architects do have
44:24
accountability over the entire portfolio of tools that align to that product some of which that may not be actually owned
44:30
by it because those products are still performing a function that that business needs so the solution architect needs to be
44:36
aware of that in terms of sort of stitching that broader context of say a marketing product or a sales i.t product
44:42
together to get that full picture and just oh actually let's take another question
44:49
i was going to just say there's the build component too right so if they're going to build it versus buy
44:55
it like everyone's talking about buying it but if you're going to build it provide the standard automation or whatever tooling that they
45:02
can use to create that technology and build the governance into the platform so then when that build gets deployed
45:09
it's automatically registered in compass the security scanning happens on all of the libraries you have a trust score and
45:15
everything gets automated and it's been done in the background real quick part b question um with
45:22
respect to innovation and keeping the lights on um oftentimes i've struggled with teams trying to do both and
45:30
not able to focus enough time on innovation i mean i love the idea of you know if you're spending 80 20 you know take 10 out of it and then focus on
45:37
innovation and automation right sounds good but in reality uh just switching hats to be able to support different
45:44
systems with the proliferation of the number of tools any it companies or any id organization has to support these
45:49
days it just has become more and more harder with same group of people trying to quickly be able to keep the lights on
45:57
and also do innovation do you have any recommendations on whether there's going to be separate dedicated teams that
46:02
would benefit from being able to innovate more quickly or so we had the separate team
46:10
but there are so many different domains and specialties in it now
46:16
so that team is very good at looking at you know the bigger trends and they look at you know things that are
46:23
down the second or third horizon like virtual reality and blockchain and potential use cases within atlassian but
46:29
there's still a lot of innovation to be had in say hr tech and sales tech and finance
46:35
technology and because they're doing it day in day out they will be able to come up with better
46:41
ways of working and better ideas for their particular domains and that's why we need to encourage that
46:47
decentralization of innovation thank you yeah um i think we're gonna get kicked out of
46:52
the room if i don't stop taking questions um but here is the qr code if you want to take a picture of that to our microsite where
46:58
we keep we've been publishing content it's a new microsite i think it's just been up for a few months but we're going to have more content coming out of there
47:04
as well as um hopefully some links to some of the presentations that were here this week for it but thank you so much
47:11
appreciate your time and great questions hope you have a great rest your day [Applause]