Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Building Data Driven Organizations
Search
Abe Stanway
September 13, 2014
Programming
1
210
Building Data Driven Organizations
Given at IT Weekend 2014 in Kiev
Abe Stanway
September 13, 2014
Tweet
Share
More Decks by Abe Stanway
See All by Abe Stanway
MOM! My algorithms SUCK
astanway
15
2.8k
A Deep Dive into Monitoring with Skyline
astanway
6
1.8k
Bring the Noise: Continuously Deploying Under a Hailstorm of Metrics
astanway
34
7.9k
Data Visualization in the Trenches
astanway
5
710
Gifs as Language
astanway
2
830
Your API is a Product
astanway
3
970
Zen and the Art of Writing Commit Logs
astanway
3
820
Other Decks in Programming
See All in Programming
すべてのコンテキストを、 ユーザー価値に変える
applism118
2
420
データの民主化を支える、透明性のあるデータ利活用への挑戦 2025-06-25 Database Engineering Meetup#7
y_ken
0
280
無関心の谷
kanayannet
0
180
Passkeys for Java Developers
ynojima
3
880
Create a website using Spatial Web
akkeylab
0
290
エラーって何種類あるの?
kajitack
5
270
今ならAmazon ECSのサービス間通信をどう選ぶか / Selection of ECS Interservice Communication 2025
tkikuc
11
2.8k
「ElixirでIoT!!」のこれまでとこれから
takasehideki
0
370
赤裸々に公開。 TSKaigiのオフシーズン
takezoux2
0
140
A2A プロトコルを試してみる
azukiazusa1
2
810
WindowInsetsだってテストしたい
ryunen344
1
190
Effect の双対、Coeffect
yukikurage
5
1.4k
Featured
See All Featured
Making Projects Easy
brettharned
116
6.3k
The Power of CSS Pseudo Elements
geoffreycrofte
77
5.8k
Why Our Code Smells
bkeepers
PRO
337
57k
How to train your dragon (web standard)
notwaldorf
92
6.1k
Thoughts on Productivity
jonyablonski
69
4.7k
Gamification - CAS2011
davidbonilla
81
5.3k
The Pragmatic Product Professional
lauravandoore
35
6.7k
Imperfection Machines: The Place of Print at Facebook
scottboms
267
13k
Raft: Consensus for Rubyists
vanstee
140
7k
The Web Performance Landscape in 2024 [PerfNow 2024]
tammyeverts
8
660
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
31
1.2k
Balancing Empowerment & Direction
lara
1
340
Transcript
@AbeStanway BUILDING A DATA DRIVEN ORGANIZATION
1. why 2. how
1. why 2. how
“DATA IS THE NEW GOLD”
Predict the future!
Retain Customers!
Grow the business!
Recommend content!
Drive Engagement!
unclear paths to $$$
IN IT, It’s clear.
Data are Dollars
. IT Working = +$$$ IT Not Working = -$$$
. .
How do you know if your IT is working right
now?
How do you know if you are earning money right
now?
KPIs. What are they?
Etsy: Literally a Money per second Graph
Planet Labs: Literally an Images per day graph
What are the Kpis for kips?
$ per second items bought per second page requests per
second database queries per second memcache hits per second fread() per second
If you do not have the data about your infrastructure,
it is already broken. LEsson:
None
Test driven development -> data driven development
Without data, you are flying blind
How do you know you’re hitting your goals?
How do you know if You’re making the right ones
in the first place?
How do you know if you’re still in business?
How do you even know what planet you live on?
Assumptions are death
You need data, yo.
1. why 2. how
1. collect 2. analyze 3. ??? 4. Profit!
1. collect 2. analyze 3. ACT 4. Profit!
data that cannot be acted upon should not be analyzed.
None
You are running a business, not an art museum
You are Trying to Win the market, not a fields
medal
This can be disappointing
Data SCientist?
Data Scientist? Realist.
Find a way to Align your employees intellectual curiosity With
your Real business needs. LEssoN:
Train your organization
you need a data culture.
“It’s not shipped until it’s monitored”
“If you are not looking at Dashboards, you are not
doing your job”
Building instrumentation and watching dashboards are hard And Time consuming
App code -> statsD -> Graphite -> Dashboards -> Insights
by hand by hand by hand by hand by hand
Developers just want to code
Let’s automate
Which is easier to automate? Insights or data collection?
Insights are sexy and fun
Collection is hard And unsexy
Collection is hard And Boring
Collection is hard And unsexy
We’re on track to have excellent automated insights
anomaly detection
App code -> statsD -> Graphite -> Dashboards -> Insights
by hand by hand by hand AUTOMATIC! AUTOMATIC!
(…if only we had the data)
How do we automate data collection?
currently have ganglia, New relic, collectD, etc
NOT WHAT WE NEED
they provide data about your raw machines, not your CUSTOM
DEVELOPED TECHNOLOGY And Application level logic
Healthy servers don’t make you money. Healthy services do.
enter LARIMAR
Full disclosure: this is my new PROJECT ! we’re going
to talk about it because i’m pretty excited and the beta is opening up soon.
LARIMAR uses raw machine metrics to infer App level architecture
and inform developers about problems
A service: cpu resources disk io PCAP data ports Used
syscalls
A service: cpu resources disk io PCAP data ports Used
syscalls service fingerprint MACHINE LEARNING
A service: cpu resources disk io PCAP data ports Used
syscalls ABNORMAL BEHAVIOR MACHINE LEARNING
a system: service service service service service MACHINE LEARNING graphical
system fingerprint
a system: service service service service service MACHINE LEARNING Abnormal,
holistic system behavior
Larimar automates both analysis And Relevant data collection
so your developers can focus on coding and acting on
insights
No configuration!
App code -> statsD -> Graphite -> Dashboards -> Insights
by hand AUTOMATIC! AUTOMATIC! AUTOMATIC! AUTOMATIC!
1. collect 2. analyze 3. ACT 4. Profit!
organizational shifts are still needed to inspire ACTION on Data
but ACTION is easier to inspire when there is lots
of data and lots of insight everywhere
Create a culture where your developers create these kinds of
tools
When a data driven mindset is the default, tools will
build themselves.
Thanks! @abestanway ! ! larimar.io @larimarhq