- [Lisa] All right, folks, we are about ready to get started. As we go, if you've got questions, or you need to switch your audio, you have those options in that sidebar, but I wanna welcome you to our session. And thank you so much
for being here today, to learn about Getting Started with Cloud Financial Management. I am Lisa Harnett, the host for CFM Talks, and these talks are
about 30 minutes in time, the rest of is devoted
to questions from you. So as we design each of these talks, we design
them to answer
questions and common questions that we hear from our customers. So today's talk, we are talking about people
who are starting to get set up with Cloud Financial Management programs. Are you asking about allocating costs, how to make people accountable, how to stay on budget? Or starting to figure out whether your different applications are profitable, not profitable. How do you start
answering these questions? That is what today's talk is about. So as we started this call, we had
a poll asking you what
your experience in this space, you know, what's your role in this space? And so I'm gonna go
ahead and share that back to give you an idea where, what we have in the call today. So 44% of you guys are
not in any kind of a role related to Cloud Financial Management, and the rest of you,
most of the rest of you have got less than one year, but there is also a sizeable
chunk I see out there with more than two years. So some of this will be a
reminder for those of you guys in
the more experienced range, but I think this is a great place to be if you're just getting started, so thank you so much for being here. And know that you're not alone. We've got a variety of experiences here, and the link for this presentation will come out to you tomorrow, and an email along with the recording, if you wanna go ahead and share it out. But at this point, I'm gonna
go ahead and turn it over to Lauren and Bryce who will take you through the agenda. Thank you so much for being her
e. And as I mentioned before, go ahead and put those questions
in that question panel. And we've got lots of time in this to hold out for answering them, thanks. - [Lauren] Thanks Lisa. Hello everyone, and
thanks for joining today. We are gonna be walking through, getting started with Cloud
Financial Management. So the first section
we'll be going through the different concepts that make up CFM. Once we all have a common
understanding of that, we'll then be diving into
how do you actually take t
hose concepts and turn
them into an action plan. So what steps do I need to go through to actually start my journey? And then we'll wrap up we'll
with what tools does AWS have to help you along the way? So starting with our concepts, if you've joined one
of our webinars before, you're probably familiar with the screen. We break up our pillars of CFM into these different phases here that you'll see, Plan, Run, See, Save. One thing to note is that this is not
chronological by any means. So you mig
ht have seen this
in a different arrangement of See, Save, Plan, Run, but we have listed below
each all of the tools that AWS has that can
help you along the way, in each of those four pillars. So while these are concepts,
today's focus is gonna be on, how do I take these concepts, and again, actually have an action plan? How do I start that in
my own organization? And so we'll be going through
a six-step plan for you today, of how to actually get started. First thing to note here is that the ne
w consumption model of cloud is very different than
your typical data center. So you'll notice here that
you're bringing together technology and finance teams, and technology teams aren't
just technology teams anymore. They're doing things that
are normally associated with finance or operational
teams, they're product owners. And so air role takes on a
little bit of a different shape with this new cloud technology model. So knowing that this is so different than what you might be
used to in the
past, how do we set you up for success? We'll walk through what tools we have, and again, that six-step plan to make sure that you're successful across this journey, no matter where you may
fall on that spectrum of, I just got started, I'm halfway through, or you wanna iterate upon
what you already have going in your organization. So where I typically like
to start when first engaging with customers is asking this question of, what does good look like to you? Very simple question. And oftentimes
that answer
can be either unknown, or really complex. So this graph kind of helps
demonstrate a couple things that can help you answer that question. First off, when you're
looking at this graph, our focus is on that purple
line, that unit cost. Not so much on those teal bars
that are representing usage. That unit cost can be something
like cost per customer, cost per transaction, number
of customers that you have. It's a business metric
that makes sense to you and your organization. So by focu
sing in on a KPI,
a key performance indicator, like that unit cost metric, we're not so focused on that usage. You might see your usage scale up, but there's always gonna
be some optimization lever that you can pull along the way that can help you to
reduce that unit cost, to show that you're being
more efficient in your usage, even as those teal bars may grow. So yes, your bill might
look a little different, it might be growing, but if you're seeing
that unit cost go down that purple line, you'
re
able to do more with less. So by focusing in on that, we can really help set you up for success. Once you figure out what is
that purple line for you? What unit cost is really attractive to me, or can help me to measure
success in my own organization? Once we have that, we can
start on this six-step plan that I've been referring to. So the first step here is CFM ownership. Determining a person or a team
who's ultimately responsible for all of your CFM activities. This will really help
you whe
n you're driving any sort of execution of, we wanna start tracking X, Y, and Z, we wanna report this way. By having a team or a
person dedicated to that, you lose that finger pointing of maybe someone else is
doing cost optimization, so I don't have to. You really get that
ownership piece set first, so that the teams know who to go to when, if they have a question on CFM, and you can start driving
change within your organization. Once you have that piece in place, you really wanna make sure that
your finance and
technology stakeholders are best friends throughout this process. They should constantly be in line, and that organizational
partnership will help solidify what your work is doing. By having a common language
between those two teams, they might not be focusing
on the same things on their day-to-day basis, but
in terms of CFM activities, you wanna make sure that
they're aligned in everything that they do, so that moving forward, you have that input from both teams. You wanna mak
e sure that
they're able to participate in each other's activities, such as having the technology folks in on those cloud budgeting
meetings with finance. And then finance, maybe learning
from the technology teams on different payment options, what spot usage might look like. So having them become more cloud savvy while then the tech teams
become more finance savvy will help to really
drive forward that change within the org to I'm
a cost to wear culture. Those first two steps are really those f
oundational building blocks. So those are gonna be the most important to get set in stone right away, so then you can focus
on steps three to six, which AWS has tools to help you with. Those first two are really strategy based, and more of the softer
skill side of things. So once you get to this third part, this enabling cost allocation, that's where we can bring in
some tooling from AWS's side to help you with that piece. So being able to tie back the cloud usage to the lines of business that
a
re ultimately responsible for that usage is key here. And you can use things like
AWS Cost allocation tags to help you along the way to be able to do that and
to report on that easier. Once you're able to do that, you wanna make sure that those
folks who are responsible for the cloud spend have
visibility into that. So things like AWS Cost Explorer can be extremely helpful there, by providing those dashboards that'll be unified across teams. So if you do find those KPIs
that you're interested it
in, if you are asking for
change within a team, or a cost center, they can see what data
you're referring to, and it's the same across the board. So being able to visualize
your spend is really important, and it'll help with that KPI aspect. Then we can move on to that
cost optimization piece. What levers do I have in my
back pocket that I can pull, so that I can consistently
be showing more efficiency in my usage, doing more with less. Here's where things
like AWS Cost Explorer, EC2 Rightsizin
g
Recommendations are fantastic. This will pinpoint for
you different instances that could either be terminated if they seem like they might be idle, or that there might be
a Rightsizing option that you can downsize, to then be more efficient
in your EC2 usage. By having this pinpointed for you, you don't have to go search
for those opportunities, AWS will bring those to you, so it reduces on that
manual section of reporting. And lastly, in the six-step plan, we have our Forecasting
and Monitori
ng piece. No one likes that surprise bill. So how do I make sure I
don't have that at the end of every month? By using things like AWS Budgets and AWS Cost Anomaly Detection, you can avoid that surprise
bill by being alerted when you might be just close
to reaching that threshold, or when you've actually
reached that threshold. So you can do that based
on forecasted spend, or actual spend, so that
you're not surprised at the end of the month,
and you can take action, and be more proactive, if yo
u are budding up
against those thresholds, you've set in place. And the last piece here,
not necessarily a step, but something that should be integrated throughout your whole cloud
journey is continuous education. You wanna make sure that your teams are consistently improving their cloud IQ, whether that's through doing something like the AWS Cloud Practitioner Exam. You can take that virtually now. So that's one way that you
can continuously improve your cloud IQ, or attending
things like works
hops, or reaching out to AWS, we can help in a variety
of different ways. But always being able to
improve your cloud IQ, making sure that you're all
on the same page as the cloud is ever evolving, and we're
always introducing new services and new tools that can help
you along that journey. So now we've gone through
that concept side of things. How do you actually move
this into an actionable plan? I will hand this over to Bryce
and he'll walk through then all of those steps in a
little bit more
detail, on how you can actually start implementing those into your org. - [Bryce] All right, thank you, Lauren. One thing that, for those
that may have joined late, that we do encourage is questions. We certainly have prepared time at the end to make sure we get questions addressed, but if questions come up for you, please feel more than welcome
to put them in the chat now, and we can make sure we benchmark those, and grab those for questions
later, appreciate it. So as we start our journey, a
lot of customers that we
engage with are pretty surprised that what comes up first is
about this idea of ownership. They're surprised that it's not, well, what you need to do is
you need to go get wins. And so you need to
establish a secret practice to drive efficiency in the portfolio. Or it's not starting with
visibility to say that, "Well, I don't even know where
I'm spending money today, "so let's go work on
the visibility piece." Based on experience, the reason why we start
with ownership i
s because even in those instances where
you need to drive efficiency in the portfolio, and you're gonna be interacting
with engineering teams, as soon as you start to try to
build a partnership with them to say, "Hey, we notice
that these EC2 instances "can be right size," or,
"We recognize that there's "an opportunity for Savings
Plans to be purchased." The response back will be, well, what authority
do you come to us with? Why should we pay attention? And that's where this
ownership piece beco
mes very, very key because upfront, if you're able to establish the importance of the CFM practice within a team, within an organization, or more
broadly across the portfolio of the entire company, that's
then provides you the mantra. It establishes that strategic
vision that you can set in that mission statement about what is your team tasked with, who's responsible for tracking
and maintaining costs. And then you can then
go, later, when we get into the later steps
and talk about building thos
e organizational partnerships. Because in this ownership piece, well, once this is established, that gives you that opportunity to know, I am responsible for these types of costs. And I will start in, not only getting the
visibility into the portfolio, but also helping drive efficiency. And so, as we talk about
the actual implementation of this ownership piece, this is really what starts
that unfreezing process to help migrate, remove the organization in the direction of cloud financial manageme
nt, because you're creating that
need within the organization, that visibility that's needed to then go start building
those partnerships, which is that second step for us. So in that partnership piece, the reason why this comes up second, is because once we have
that ownership defined, we have to recognize that
our CFM journey collectively, as an organization, is
not going to be complete, unless we bring others on
the journey with us, right? So it may be that you
started out your CFM journey be
cause tough questions
were getting asked about, well, who's spending the money,
where are they spending it? And what are they spending it on? And that may have been
from the finance people, or it may have been from
the engineering side where they say, "Hey, you know, gee-whiz, we're spending way and more on our product than what we had originally anticipated. Why, what's going on? Why is that? And so it's in partnering with these different organizations, you start to build a
robust practice that
helps to really understand the visibility side, because as the organization
starts to implement a CFM journey, there's going to be items
within the organization that also have to be updated, and historically may not have
had to have been addressed. For example, in working with the finance organization, internally, you may need to start update updating your different internal
accounting practices. You know, historically when
you engage with a third party to provide services, you probably did it
at a negotiated rate for a set contract period. Well now in the cloud where
you have variable spend being introduced on a large scale, the finance may be saying,
"Hey, we have a problem," because the purchase acquisition
that was established, isn't matching the month-end bill to be able to pay back that invoice. So, because it doesn't match, there's nothing that we can do, right? And so that's where it starts
building this partnership with the finance organization, to really start understanding
this Cloud Financial Management, and how the organization needs
to be moved in the direction of maturing in that practice. With the engineering side of the house, you know, what's funny is
I've never met a engineer that enjoyed spending
money in the cup, right? They were always trying to
produce some type of value, some type of function. So as you start to engage age
with them to do Rightsizing, or to introduce efficiency
into the portfolio with maybe adoption of new
functionality coming from, l
ike, for example, GP
three, for your EBS volumes, if you have that established partnership with those technical teams, you are able to understand the
potential impact of decisions that are being made. And you can know start avoiding situations where you're trying to be penny wise, but pound foolish in trying to implement a cost optimize mechanism
because an engineering team, as a partner was able
to influence and guide that CFM journey. So those are some instances
there of making sure that that
partnership
is being established, and how you go about doing that. Once we have those first
two to opportunities with the ownership and
the partnership defiant, that's where we get into the third step of really understanding
that cost allocation piece. This in like what Lauren
mentioned at the beginning, this is where we could
start introducing tooling to really help accelerate that journey. And the previous two steps, it's very much trying to influence without necessary authority,
or explicit o
wnership. Even if it's been buy-in
from the highest levels about CFM and the need
for those practices, it could be challenging
to then start interacting and engaging those partnerships. And so the reason why cost
allocation then follows on from those first two steps is because unless someone
is getting the bill for the charges that they're incurring, it may be challenging to get
buy-in into the CFM program. You know, in referencing
back to the engineering side, they also, you know, even
in the s
urvey we saw most, or maybe not over 50%, but pretty close to 50%
of the folks on the call are not in a traditional CFM position. So you may be the recipient of a lot of what the CFM
organization's coming to you, and talking to you about. So you're living this
sort of day in and day out that you're not enjoying the inefficiency that's being spent in the cloud, but you're challenged
with how do you balance what the CFM team is telling you with delivering value in the cloud? Because that's what yo
u're
being measured against is, can I deliver a new capability
or functionality in the cloud for the application or
workload that I'm supporting? And so then you have the CFM
team coming to you saying, "Well, you know, the boss over here says "that we need to go do this on doing that, "as part of the ownership, "we need to build a partnership." And you're like, "Yeah,
but how do I ensure "that this is being prioritized?" That's where the allocation
piece starts to play into it. If we can ensure
that your
costs that are being incurred are flowed back to your bill, then we can help understand
where that priority lies. Are you facing budget constraints in being able to deploy that
new function work feature. Because if you are, great, this is where CFM starts to come into play because the CFM individuals
can help you drive efficiency in the portfolio that
creates room in the budget, that then allows you to ensure
that at that original focus on driving functional
performance is carried out
as part of your strategic goals. So that cost allocation piece, if you can partner with the CFM folks, or conversely, if the CFM individuals that are starting out on their journey can identify the right
teams and get the bill flowing back to the right teams, that helps generate a lot of reason to make sure that folks are
being brought to the table from a cost financial
management perspective. All right, and so no
matter where you're at on your journey, right, there's going to be a lot of
differe
nt tools to help you. So if you're at the very low
end of this spectrum here, let start, let's go out there. Let's get a copy of the invoice
and the billing consult. Let's take a look at it, and start getting the first
picture of where these costs are start going to, because maybe that's where you're at. You've defined the ownership. You understand it's important, or maybe you've seen a copy
of the bill and you realize this (chuckles) this
needs to be important. So you've started, you've
got tha
t ownership in place, but you're looking at, where do I start building
these partnerships? Let's go back to the invoice. Let's let the data lead us to that answer. And that's where you can
start identifying those teams to start working with. And as you start understanding
that billing mechanism, then we can go into cost export, to start establishing the
cost allocation and saying, "Oh, well, now that we're
starting to tag stuff, or now that we're starting
to have cost associated to a particular
linked account, we can then have those costs start to flow back to the appropriate teams." And as the organization matures, you're gonna have a lot
tougher questions being asked. That's where that detail
that's provided in the cost and usage report, which is on the very
far end of that spectrum provides that dexterity to be able to get the right reporting for your organization, to help meet those
little unique challenges that start to really bubble up, as your CFM journey
starts to really take o
ff. And there's also third party tools, so, you know, again, don't
try to go this alone. What was mentioned at the very
early part of this call is, you know, it takes a very
wide ranging set of activities across a broad spectrum of tools to really make sure that your
CFM journey is progressing as it needs to go. And that includes working with
very trusted third parties through the AWS platform. So please keep those individuals
and those teams in mind, as you get to these
points, and you're like,
"Hey, this may be a roadblock "that we may not be able
to solve internally." So on the next part, what we're gonna do is pass it back to Lauren. And she's gonna talk a little
bit about the KPI structure, and why that's very, very important to this cost visibility piece, is step number four. - [Lauren] Thanks, Bryce. So the second and half of
that cost visibility journey is establishing those efficiency
KPIs in your organization. Going back to that, what
does good look like slide, looking at tha
t purple line, what do I care about? What's my unit cost? So figured it would be
helpful to throw in here what the heck is the KPI, and
why should I care about it? So KPIs are measurable. They're meaningful to your business, and they can help you to
measure what success looks like. So, one way to do that is work backwards from what you're trying to do, right, work backwards from that desired outcome. And how will I know if
I'm being successful? These can be things like
I'd mentioned before, cost
per customer cost, per transaction. Things like that, that you can say, "I'm working towards X goal," a KPI will help you get there. And they're really important in that organizational partnership piece. So yes, they provide great
visibility into your data, but they also help to align stakeholders. And maybe they don't have a
common language day to day, they're focused on two
very different things. But those KPIs will say, these are those top priorities
that we're working towards, right, this i
s what we
consider are good to look like in our organization, so this
is what we're striving towards. This will help you to set those targets, and then it'll also help you
to see are my CFM activities that I'm doing on a day-to-day
basis, improving this metric? And then if not, what
levers do I have to pull so that we'll be driven towards our goal. So KPIs can help you on
the visibility front, and then the organizational
partnership front as well, to see your efficiency metrics over time. Now th
at we have those
building blocks in place, we can go forward to the tools
that AWS has to offer you to help along the way. By no means am I going
through these one by one, I promise you, but this is a slide of all of the different
tools that we have available that'll help you with your CFM journey. The purpose of this is
just to show that there's a wide range available, and we're constantly adding to this list. So as that's evolving, you wanna make sure that you're cloud IQ, like we talked about
before,
you're constantly evolving. You're checking out if
there's something new that'll help you in your journey, and then exploring different options. If you are really liking one, you can stick with that. But if something else isn't
quite working for you, there's a wide range of tools
available for you to have. Getting more into our latter
half of that six-step plan on step five, we're talking about Cost Optimization. So while there's hundreds
of different levers that you can pull, I have li
sted out here
three different ways that we can help you with AWS tooling to optimize and save. The first two here, Savings
Plans Recommendations and Reserved Instance Recommendations, through AWS Cost Explorer,
these will be presented to you. So you can toggle between your date range that you wanna look back on your usage for your commitment term. Are you looking at doing a
one year versus a three year, and how do I wanna pay for this? So they're a commitment based. Do I wanna pay all upfront,
p
artial, or no upfront? You can toggle between those choices to then get that customized
recommendation for you. What this is looking
for is where can I snap that chalk line in my usage to say, "This has been consistent
over X period of time. "I know why I'm always gonna
be using at least this amount. "I wanna save on that usage if I can, "'cause that's consistent." So you don't have to go
find that opportunity. We have tools that can
help you to pinpoint those, and you don't have to
search for t
he results, they come right to you. That will help decrease
that manual reporting piece. And it'll help you be more efficient in your usage and your time. On the right hand side, we have our Amazon Elastic Compute Cloud Rightsizing Recommendations. Also available through AWS Cost Explorer. And that's gonna help you
with your EC2 efficiency. So that's gonna identify things for you like Idle Instances or
Rightsizing Recommendations. So between one and 40% CPU utilization, what do I have of instanc
e wise that lies between those two metrics? And what are my downsizing opportunities? That tool will also show you
your estimated monthly savings, so you can start to prioritize. Again, I think the whole
theme on this slide is, we do that work for you. These tools are working for you, so you don't have to manually
go find those opportunities they're presented to you automatically. And as we get into our last step, the Cost Forecasting and Monitoring. We have tools like AWS Cost
Explorer that can
help you with tracking your KPIs, having that baseline of
where you're starting is always helpful from when
you're looking to forecast. How do I know what good looks like? How do I know what my normal spend is? By using visibility
tools like Cost Explorer, that can help you get your baseline, then you can forecast from there. So with this forecasting function, it'll help you plan out what
your future could look like, and give you thresholds to you, so you can know what the
high end and the low
end of that forecast could look like. As we move forward to the budget section, AWS Budgets has alerts and
actions that you can take, so, as I had mentioned earlier on, you can be alerted when
something is forecasted to go over your budget, or when it has actually
reached your budget. That's really important
to have a mix of those, so that you can be as
proactive as possible in this variable spend environment, right? So you wanna make sure that
you have those alerts set up, so that you can const
antly
be taking actions to minimize that surprise,
at the end of the month. You'll know what you're spending. With those AWS budget reports, that'll help you to see where are you in terms of your actual budget? So what are my actual versus goal? And those are available to
you in a dashboard format. So you can constantly revisit those, and tailor, if you need
to make any changes. Next steps here then are
really all revolving around these six steps. Where are you at in your journey? So whether you
're just starting out, you're halfway through, or you have a plan of your own, using these six steps to
then guide you and say, "Okay, I really need to focus
in on these top two," right? Bryce and I both mentioned
those top two priority areas, are really huge to have in place to make sure that you
have a solid foundation for the rest of your CFM work. So use this as a playbook
to just start iterate upon and really invest in your CFM journey, all of these will help
set you up for success. And als
o know that AWS
is always here for you. We have a wide range of different
things that we can offer, such as our hands on workshops
with your account teams. We have CFM peer connects that
you can reach out to others that are in their own Cloud
Financial Management journey, and see how they've defined success, and what steps they've come across. We have a wide range of things available, so feel free to reach out to
your account team after this, to get started on that. And with that, we will open
u
p for our Q&A session. - [Lisa] Thank you so much, guys. Sorry about that. I will go ahead and start
with one of the questions we have out here, which is, how do we get started on the CFM implementation side? So where would you start
for the implementation side? I think this question came in early, but it would be, you know, after you get past that ownership piece, where would your first step be for that? - [Bryce] Yeah, absolutely. After we get the ownership piece defined, and we start understa
nding a
little bit about the portfolio, that's where I would start, is let's go out, let's
take a look at the bill, and let's start understanding
a little bit about the operating environment. What type of applications
do we have operating in the cloud today? What type of value is
this being delivered? You know, what is the
context that you can bring to make it context aware? When you're looking at
the cost and seeing, oh, okay, now understand
why we're using EC2, or understand why S3,
or RDS is
being used. Because it's in support
of this broader strategy to deliver functions to the organization. What about you, Lauren? How would you answer that? - [Lauren] I think you hit
the nail on the head there. I think that organizational
partnership piece is all about cross education, right? You have these finance
teams and technology teams that might not be of in the
same day-to-day activities. So that cross organizational
understanding, whether that's through KPIs
or just educational offerings,
making sure they understand
each other's worlds is huge. So if you come to a budgeting
session and you're like, "I don't even know what EC2 is." That might be really hard
to start budgeting for that. So making sure that you both
have that common understanding will help make that a stronger
organizational partnership, and then lead you through
those like three to six steps where those tools then come into place. - [Lisa] Okay, and so, I
have a follow up question. I don't have many services
yet d
eployed on AWS. What can I use to proactively
build with a budget in mind? So I think this is gonna
be the pricing calculator, or depending on how big your migration is. Is there any other
ideas you have for that? - [Bryce] I would definitely
leverage the AWS calculator for that, for sure, 'cause you can build that out
on a sort of workload basis, if that's what's appropriate to you. But you can also take a look
holistically at the portfolio to say, "Generally speaking,
if we need to deploy "X a
mount of compute into the environment, "what is that gonna look like?" So it's pretty flexible
in it's capabilities. - [Lisa] Okay, and then the other question was along those same beginner lines is, do any of this cost
visibility optimization and monitoring tools have a
cost associated with them. - [Bryce] All right, Lauren, take a stab 'cause I know in general they don't, but I know that you have
a lot of work with those. - [Lauren] There's always an
exception to the rule, right? So in general
, the answer is no. A lot of these are available
for you free of charge. If you do get more customized
and start using things a little bit more advanced, as Bryce was talking
through that matrix slide, of your visibility options. As you climb that ladder there, that's where you would
start seeing those charges when you start using things
like a third party billing tool, or you start using QuickSight dashboards, things are on the more mature side, that's where you'll start
seeing those charges co
me in. - [Bryce] Yeah, so some good
examples, like Cost Explorer's not going to cost you anything. - Unless you turn on
- And there's some- - [Lisa] hourly granularity? - [Bryce] Yeah, exactly, and that's where- (chuckles)
- It might cost you a dollar. (Lisa chuckles) - [Bryce] (chuckles) Yeah. - [Lisa] Depends on how many
things you have. (laughs) - [Bryce] Absolutely, and
very much to Lauren's point, it's really the extension capabilities that you get in there,
but natively, right? So for exam
ple,
- Mm-hmm. - [Bryce] Compute Optimizer, and that's gonna come up
in one of these questions about EBS volumes, when we
talk about the optimizer, that's a free tool. But if you start leveraging
some of the latest capabilities that were announced at
re:Invent most recently with being able to extend metric data to the 90-day mark, that's where I'm sort on a per resource, or account, or organizational level that you would then optionally
start to encrypt costs. Anomal detection is free. AWS Budge
ts is free, but if you then start
using some of the actions, and what those actions are carrying out, as part of the budgets, those would start incurring costs. So hopefully what what's
being illustrated here is at a base level the tools will be free but as you start to extend
them you may incur costs, depending on how you leverage
them in your organization. - [Lisa] Yep, that makes sense. All right. So did you wanna take
the EBS question then? How do I find out the utilization of EBS volume?
(B
ryce laughs) Since you mentioned it? (Bryce chuckles)
(Lisa laughs) - [Bryce] I feel like
maybe the product team for Compute Optimizer
wrote that question in here so they could glance.
(Lisa laughs) That is the exact use case
for Compute Optimizer. Yeah, and so Compute
Optimizer will support four different services. It'll do your EC2
Volumes, or EC2 Instances. Let's just say your EBS Volumes,
your auto scaling groups, and then your Lambda functions. It will cover those four services. As it relat
es to EBS Volumes, the beauty of the recommendations that are gonna be provided there is that it's gonna take a look at any restriction deployments, and tell you, "Hey these,
or even if it is optimized, "here's where you can still save money." Good example is any gp2 volumes. Even if it is the right storage size, and you're utilizing it appropriately, as it relates to throughput and so forth, you can still save money
potentially by going to gp3, and that tool outlines those. And in line, in that
recommendation for that volume, it will show you what is the performance of that volume over time? I think default is 14 days. And if you were to select, you know, hey, move to gp3, it'll show you what that
marginal effect would have been had gp3 been in effect for that volume over that period of time. And sufficed to say, it'll do that across other recommendations as
well within EBS volumes. - [Lisa] All right, I'm gonna go ahead and transition from that over to, we've got a lot of
questions a
round tagging. So talking about tagging data quality, setting up an appropriate
tagging structure, what is your advice on, you
know, kind of monitoring or setting up your tagging? - [Bryce] All right, Lauren, you first. (Lisa laughs)
- I'll start. I think tagging is awesome, right? But you don't wanna get to
that point where you have so many tags that now they're
all invaluable, right? So I would pick what's
really important to you. Do I wanna know what
project, what cost center, what person or
team this belongs to. Start small and you can expand upon those. But once you get to the point
where everything's tagged with 300 tags, are those
really valuable to you? So I would start thinking through of that start small approach
and then expand as needed, but having those core
tags will really help you to gain better insights from your data. - Mm-hmm.
- Absolutely. And that's the, you know, more over to
Lauren's point on that, that start simple, try to identify those
opportunities where you
can put one tag on the resource
that then can link out to a broader data set. And the reason why you reference that, for example, an app code, or some type of unique
identifier for the application. If you could tag the resource, just the app code, but then in a separate
data source reference how many users is that application? What type of risk assessment
is that particular application? How old, or what is the retirement plan for that application? If you can do that, that will avoid you having
a
tag on the information, on the resources, and as the
organization experiences churn, or different org unit changes
that happen all the time, by having that app code there, you could change your data
repository as much as you want, without having to affect that ID code for that particular application. So that's a good example of, are there different mechanisms you can use to minimize that tag count? - [Lisa] Yep, so then what about tracking kind of that data quality? Like how do you, is there an
y tools or services that we have for helping with that? - [Bryce] Absolutely, so the
first one that you can start to reference is tag editor. That's a good tool to be able to go out, and view at a per resource
level potentially, which items have received the
tag that you're looking for, or haven't, and it allows you to sort of also view the
iterations of those potential tags. So for instance, if you
have a tag key value pair such that you're gonna, in referencing back to that app ID, let's say t
hat's required. And so you have a key of app code, but you'll start to
notice in the tag editor that the enumeration of the value is such that it's starting to skew beyond what you need from a data quality control. That's where tag editor can come in and start to help you
identify those settings. You also have AWS Config where
you can actually set a policy that says, it has to be
tagged in these certain ways before a resource can be deployed. It can also measure drift
from a known good standard.
AWS can do that, AWS
Config can do that for you. And I'll also will reference a third tool, which is your curve file,
your Cost and Usage Report. If you have taken that tag, and you've enabled it
as a cost allocation tag in the billing console, what you will be able to do then is see in the Cost and Usage Report all of the resources
that have been tagged, and what are the associated values. And then you can start to measure, are you moving in the right direction? And because it's in the curve f
ile, it now will also flow
through to cost explore, where you can group by and filter based on those tag key and value pairs. - [Lisa] Yep, from a cross control method, I'm gonna add, I'm gonna stop, probably to let Lauren step in, but I think there's also
another couple options. I know that in budgets, there's
now a way to set a budget for things that aren't tagged. So you can, either find
items that aren't tagged, and kind of get an idea of if
that's getting outta control, and cost categories
to help
kind of reduce some of that. If you've got the same thing named it, but sometimes it's capitalized,
sometimes it's lowercase, sometimes there's different
variations in there. You can use cost categories to kind of group those
back together again, and get some of that control back. Anything else Lauren, you wanna add? - [Lauren] The only
thing I was gonna add is that those control role
pieces can go from, hey, you probably shouldn't do that, to you actually can't do this. So you have that
wide range
(Lisa chuckles) of like, do I, how much of that guardrail
do I wanna put up? How strict do I wanna be? So you can make it fit to
what works in your org, but just know that there's a range there in terms of what you
actually can shut down. - [Lisa] Awesome. All right. We've got some questions around KPIs, and around optimization scores. (sighs loudly) I'm trying to figure out what the best way to answer some of these. So what tool would I use to
find an optimization score? What are so
me of your favorite KPIs? What if I don't have a unit
cost structure in place? Where would I start? So what are some, I guess there's a bunch that are in that kind
of same similar vein. - [Lauren] Great questions. I love KPIs. If you've ever heard me talk, you'll know, I love KPIs. I talk about them, I feel like, all day. So my favorite KPIs will start there. EC2 unit cost, storage unit cost, great efficiency metrics. Total EC2 spend divided by
the total EC2 running hours will get you your unit
cost. That's something that
you can track over time through data that you can
pull through Cost Explorer, and show that trend in your efficiency. If you're noticing that
that's creeping up, you'll look at your
optimization levers and say, "Wait a second. What can I do here?" Do I need to look at Savings Plans, or my Rightsizing opportunities? Can I leverage spot anywhere?" So knowing what optimization
levers go with that KPI is really important so that you
can know based on the trend, if it's go
od, keep going on
your path that you're going, but if it's not going the
direction you wanna see, what can you do to bring that back down? We do have a new dashboard
that's a 200 level, so a little bit more advanced that can help to automate part of that. So as you get going on your journey of doing some of this, maybe manually, that could be a part two. The KPI dashboard will have for you then all of these KPIs laid out, where you're at in each of them, and show a green reference. If you are ei
ther at or
above that KPI metric you're striving towards, and then it'll be in regular coloring. I believe it's white text, if it's not reaching that threshold yet. So that would be one way
that you can view that through your cost and usage report, but it is on the little more advanced side at that level 200. - [Lisa] Any other, what's
there Bryce before I, - So there's another question?
- try and take another question, no, go ahead. (Bryce chuckles)
(Lisa laughs) - [Bryce] No, that's good. Ther
e's a question in
here that Paul asked, it's says, what about
ownership responsibility of the business mission, you
know, with finance and IT? And to extend that hopefully that, well, why KPIs are so important is as those KPIs are being established, they should always map
back to the IT strategy that's been laid out for the organization. What you don't want are scenarios where you're tracking
efficiency, and so forth, in isolation of where the
organization is headed. So to illustrate the point,
let's say that the
organization has a stated goal to modernize the IT portfolio. And what I mean by that
is that you have taken a concerted effort to
evaluate and start deploying manage services so that your organization doesn't have to manage all
the patching, and so forth, associated with your infrastructure. Or let's say that you're trying
to adopt containerization because you want a lot of the agility, and IT capabilities that come with that. If that is the stated it
goal, then the CFM entit
y, whoever that should be, should be mapping KPIs back
to that stated IT strategy because you can see the cost. You can see, hey, over time, if I look at the different
consumption level of compute, I'm noticing that our
EC2 fleet is increasing, but as a percentage of that portfolio, you know, Lambda for instance, or Fargate consumption is not increasing. It's staying flat. Well, now we have a problem. And that KPI that was
crafted by the CFM entity back to the IT journey, broadly to the portfoli
o, is where these now symbiotic relationships help identify early on that, hey, maybe we need to
take a different track, something's not working
in our journey so far. - [Lisa] Next one I was gonna ask is around what factors does
forecasting take into account when estimating future costs? Do you guys know? It's mostly just based on your previous history, right? - Mm-hmm.
- That's right. - It's gonna-
- That's right. It's gonna be your previous consumption. - Okay.
- Now, something that Lauren me
ntioned, and for those that
didn't have the benefit, this is the second time we've
been able to get this talk. And so in the the first chat Lauren mentioned a very good point, and Lauren, I'm gonna pass it
back to you to flesh it out. But one of the things
that Lauren mentioned is, allow your forecasting, and
Lisa, you mentioned this too, allow your forecasting to be iterative. Don't expect that the first
time you do a forecast, you're gonna get it right. You're gonna get it wrong a lot before y
ou start getting it right. And start bringing in contextual
little bits to help you with the forecasting. So yes, natively through Cost Explorer, it's gonna be based on just
purely your consumption pattern of what it sees. But if you know there is a big project coming
up, don't put the blinders on, take the broad approach and understand what's going on contextually
in the environment to help you in that forecast. - Yeah.
- Great point. And I think to expand on that too, knowing what your norm is
, is important when you're
starting to forecast, right? So if you're looking at
the last three months, and you're like, I think this is kind of what my environment looks
like for the last year, you'll wanna know that. And as Bryce mentioned, iterate upon it, you'll get better as you get going on it, the more times that you do it. But having that baseline of knowing where your ebbs and flows normally are at will be helped when you're forecasting, as opposed to just, I'm gonna look back at the las
t 30 days, and think that's maybe what
I'll do the rest of the year. So definitely having that baseline, start tracking your KPIs, knowing where you're at now, to then help moving forward. - [Lisa] Yep, so similar to
that, on that, talking about it, you know, not just monthly, but so is, the question is, is there a
cost customization service that's available on a daily
basis, not in a monthly basis? And I think, I mean,
I'm not sure where that, what things are based on a
monthly, but for the mos
t part, I think everything's looking
at the most recent data, which is like the last day. Is there anything else from
the optimization perspective that we take into consideration for that? - [Bryce] The question
may be oriented around, it's either probably one or two things. And so my apologies to the
person ask the question if this isn't the right context, but one context may be how
often are the results refreshed. And the answer to that is
at least on a 24-hour cycle for at least even Compute
Optimizer, and Cost Explorer. Trusted Advisor, at a minimum,
is gonna be once a week. And then you can always go in there and refresh the results manually, to get a fresh update on what that is. Now, the other question,
maybe the second fold is, how can I restrict the
look-back period to just a day? Because I know that there
was usage in my portfolio two weeks ago, that isn't
what I need to map for. I'm looking at Rightsizing based on purely just
what happened yesterday. In that case, what I wou
ld say is, let's go back to Compute Optimizer. And in there in those recommendations, you can set that window
to a custom window, a default, I think is two weeks. But you and narrow that down, and then it will reconstruct
that recommendation based on that narrow window that you set. - [Lisa] Okay, all right. I have a question around Savings Plans, as I'm purchasing Savings
Plans, as our usage growths, I'm ending up with layers of
plans with different end dates. How do I make this administration
easier. - [Lauren] Fun question. It's a good process that
you're doing, right? Buying that baseline. That baseline's gonna shift as you grow, so adding them on top of
another is a great practice. It just, as you know, adds
to that administration piece. So tracking those end dates
can be a little bit tricky. That's where some of our
QuickSight dashboards come into play, they outline that for you, set alerts for that. It's just as you layer them, they all work in tandem with one another, but then
the operational burden increases. - [Bryce] Yeah, and absolutely. And Cost Explorer's your best friend in the utilization of
the coverage reports, it's in the utilization reports. Maybe it's the coverage and can, but it tells you that end day, so you can filter and sort, and you can always see,
hey, at the top of my list, when I look at this, what
is the most important one that I need to be looking at based on expiration date? And that'll help you as well. - [Lisa] Yeah, and I've heard customers
that have also like
standardized that I do it on a, to every other week process, so that they know it's always, I'm always reevaluating
and kind of leveling up on a regular cadence. Other customers who are bigger
might do it on a daily basis, depending on what they need to do. Other ones who are smaller, might do it on a quarterly basis to kind of get that management
to a level that they can, they can decide on what's
the right answer is. - [Bryce] Is that... - [Lisa] (clears throat) Let's see.
Are there multiple roles
that will allow folks to see different parts of their bill without impacting other accounts? So how do you structure
sort of that visibility, like how that customer culture part, what are some of the
tools that can be used to share the right information with the right people, thoughts? - So you have a couple of options depending on what type of
information you're wanting to restrict to. I would say always let's
start with Cost Explorer. - Yeah.
- Because the, that's a p
retty baseline
access that can manage. And there's a document out
there that we can follow up with that walks through the
different permission level sets for Cost Explorer, because Cost Explorer
access is also managed through the billing console access layer. So you can get very descriptive in what pieces and visibility people have with Cost Explorer. In general,
- Mm-hmm. - [Bryce] read-only access is a base layer that everyone could have or should have. And then if the concern
is not necessari
ly, if they have access to Cost Explorer, but potentially what type of
information they have access to inside of Cost Explorer. What you can do is go
into your preferences, in the billing console, and
make that selection to say, hey, I don't want people
to see certain credits, or certain other information being applied in Cost Explorer, and you can obfuscate that
data on a linked account level. - [Lisa] Yeah. Let's see, We're getting into the last
couple minutes here, so I... There's a couple qu
estions on
certifications and trainings, and what kind of things
would you recommend whether you're finance or IT,
do you have any thoughts on, have some of those resources for folks? - [Lauren] Yes, I will say one, and then Bryce, I'm
gonna kick it off to you, based on our last conversation. So the first one, I would
definitely recommend the AWS Certified Cloud Practitioner Exam. That's gonna be a really
great place to start, whether you're in finance
or the IT section, just to have that baseli
ne understanding, and it'll really help with
your billing and CFM journey. - [Bryce] Yeah, for sure. And I know that I will never
do justice as well as Lisa does on this particular topic.
(Lisa laughs) So she may have wrote the
question, I don't know. - [Lisa] I did not write... There's like three in here. I swear I didn't write it. (Lisa laughs)
- (laughs) Yeah. If you're looking for
something that I would say is independent or agnostic,
you know, cloud vendors, then FinOps Foundation, or the F
inOps organization is a really good starting place for that. They actually have a certification
that you can go through from a FinOps perspective for
cloud of financial management. Specifically speaking about AWS, it's gonna be hard to do any training and not touch on cost optimization, mainly because that is a pillar of our well architected framework, that cost optimization is. So it's gonna be hard to throw a stone in any direction and not get
some level of education on it. - [Lisa] Yeah, and
so in
the land of paid courses, there's also AWS Cloud Fund
Management for Builders and AWS Cloud for Finance. So those are two paid classroom offerings. They're multi day, I think. The Builders one is two days, and the Finance one is
three days, I think. And they kind of give
you some of that as well, from that perspective. All right, I know you guys
have access to the questions. Is there anything else
that you're seeing in here that we wanna make sure we answer? Like there's multiple account m
anagement, and governance in here, changing your case tagging strategy. With so many services available, how do you pick (chuckles) where to start to make sure you're on the right path for a good implementation? Let's do that one. - [Bryce] So I have the benefit of knowing what is a follow on activity
from this presentation. There will be a blog post
that's released that links back, and that's actually one
of the top three questions that we see from customers is, "Okay, okay, okay. Yes, I unders
tand, CFM is important. Yes. I understand that there's a path. I literally need help. Like where do I start?" I would say start at the smallest level that you have domain over, or control over. If that's a team level, start with a team, practice going through those steps of getting a manager involved
and getting their buy-in that cloud management is important, practice identifying that
manager's cost in your portfolio, practice going and
working with the engineers to ensure that they
can underst
and the cost and how to start rightsizing optimizing the portfolio, and so forth. From those experiences build on it, stretch larger in the
portfolio, stretch wider, expand that influence that
you have within your domain, or your world. - [Lisa] Yeah, and I always say, start with Cost Explorer,
it's free. (laughs) You can start there,
(Lauren chuckles) start to figure out where you, and then I always say like, what got you on this journey? Like, what was the thing
that started you off here? So w
as it, I got a scary bill, and I wanted to, and scary, and I see questions all over the place. Sometimes I see scary as a dollar. Sometimes I see scary as a bigger number, or whatever scary is to you. I got a bill, I wasn't expecting. If your first question is that, figuring out where that usage is, what led to that cost that
you weren't expecting, and what can you put in place? So we've got tools like
cost anomaly detection that can kind of help you
with those spikes or budgets that can help yo
u with seeing when things are not in control, if
that's where you need to be. Figuring out how do you set up a budget so that you can good advance notice when you're forecasted to go over what you see as reasonable. - Mm-hmm. (chuckles)
- But if you were really just wanna make sure that
people are making better choices as they make in architecture, then how do you share what
the cost of things are to back to those application teams? That's where I would start. So it's really about where, what pr
oblem are you trying to solve? Because CFM is, you can
get super complicated, but you should try and make it as simple
- (chuckles) Totally. - [Lisa] as you can to
solve the problems you have. - Mm-hmm.
- So don't add complexity just for fun. (Bryce laughs)
(Lisa laughs) I know, I-
- Yeah, thank you. - [Lisa] I totally will
go down the complex route just because I can, but,
you know, start simple, stay as simple as you can. - [Bryce] Yes, absolutely. Your future self will thank
your former self.
(chuckles) - [Lisa] Yep. (chuckles) So true. All right, last, I'm gonna ask you guys to put your last thoughts
before we go ahead and close out here today. Any last thoughts? - I can and he can-
- I have a lot. (Lauren laughs)
(Lisa laughs) - Yes.
- Go ahead, Lauren. - [Lauren] I was just gonna
say, CFM is a journey, so think of it as that, right? Everyone's gonna be at a different space, in a different spot on that journey, so just starting where you're at, figuring out what your
main pain poi
nts are, I think is a really great starting point because it can get really overwhelming. And as Lisa mentioned, it
can get really complex, so start simple, figure out where your pain points are, and go from there. And use those steps that
we had listed out today as kind of that playbook,
but make it your own. So iterate upon it, change
things as you see fit, but really keeping it simple
and starting where you're at. - [Lisa] Mm-hmm. - [Bryce] Yeah, don't be discouraged. Please, don't be discour
aged. Recognize you're not
in this journey alone. AWS, we're here to help. You know, go back to these slides, see what those resources
are for you to pull on. We had an immense amount
of attendees on this call, and what I mentioned as to represent that clearly you're not alone, and we want to help you, we want to see you be successful. So if you're starting out
and you face a setback, understand everyone does that. Even in my own professional experience, as being a Cloud Cost
manager prior to AW
S, there were nights where
we were challenged. We didn't know how we were
gonna get around a organization that was blocking us. But start small and get back
at it, as much as you can. - [Lisa] Mm-hmm. Yep. So I think that there's
one more slide in here. Lauren, I think there's a resources slide. When you guys get the deck
tomorrow and the recording that's out there, we'll
be out there as well. There are some resources that
we can point to, to help you. Even my cat wants to get involved in this c
onversation apparently. So it's a marathon. It's a marathon, not a sprint. Think about it from a, how am I gonna make decisions
based on what I'm doing. And thank you so much, Lauren and Bryce for being here today, and thank you all of you
guys for being here today. You'll get a survey as you leave, as well as an email later today, but then you'll get that,
the presentation tomorrow, and the recording will be out there, if you wanna refer back to it, but also rely on our AWS team. We're all here
to help you, so if you've got an account
team, reach out to them. There's some forums out there for asking questions as well under repost, and there's resources under our blogs, our documentation, and we're here to help. So thank you so much for being here, and I look forward to seeing
you on your next stage of the journey, thanks, guys. - Thank you.
- Bye. - [Lauren] Thank you.
Comments
Sir Excuse me, I saw your channel on YouTube. I am very interested. If you are interested in business cooperation, you can talk about the specific details in detail.