wow, it's a full house, hopefully you're not tired yet
introducing, my name Mella Camelia
I'm a Business Analyst from GO-JEK, go-pay division, to be exact
I joined GO-JEK from mid 2016
the role of business analyst is more into supporting other team
so I support marketing, operational, strategic,..
up to management level.. to know about data
my responsibility is to make sure
everyone understands about data, so they'll make a data driven strategic decision
gojek is my first job, after graduating (from university)
so, it's apt with the theme: 'kick start your data science carrier'
after my graduation, I decided that I'm going to be a Data Analyst
because I believe, with proper treatment, data is a powerful tool in decision making process
not long after I graduate, luckly I got great info
about a community, named Data Science Indonesia
they offer a Data Science Boot Camp program
from around 300 to 400 applicants
they selected less than 30 people to be trained about Data Science
basically, it's what I'm doing now on my job
so, the training helps me a lot in starting my carrier as Data Analyst
(at the boot camp) we study a lot about statistical methods and machine learning
and also (learning) a bit about clustering, regression analysis, etc
next, about data visualization, it's also important
as you know, data is numbers
it's quite challenging to understand what's the data are really about
what we can do with our data?what is the pattern?
with data visualization
with right choice and use of tools
then, business storytelling
it's about how we deliver our data analysis to our user data visualization can help the strategic decisions maker
if we don't deliver properly, the data don't have much use to the user, which are applicable to the data I'm dealingwith at my current job
last, while in Data Science Indonesia (DSI), I do networking
I met people to share& brainstorm about what Data Science actually do
and what marketing analyst social web (analyst) do
Next, about skill needed of a Data Analyst
actually, it's quite a lot
but I try to recap into these
first, numeration skill
at the job, we got big numbers of data daily
so, we have to get used with numbers and what to do with it
second, we need a great analytical thinking mind
like, what can I do with the data, what info can I give to (let's say) marketing team?
How can I help operational team so customer have a better user experience (with the app)?
third, technical and programming skills
talk about background, my background is in mathematic
we learned about the bases of programming while studying mathematic. But actually, programming is a technical skill
if you got a good logical thinking, you can learn programming by yourself
also, there are online courses you can join (to learn about programming)
and the last, business knowledge
as an analyst, we have to understand about our company's business
what is a company main goal ?
company's competitor, market research
who is considered our market
its not only internal data, but we need to find out our target market
about strategic decision
in here, I'm responsible in a team strategic
it is related with the topic
firstly, what is strategic decision ?
it's about decision but in a longterm
and tend to be future planning
what the company wants to achieve
this is one of the keys of the goals of strategic decision
and every decisions have to based on KPI (Key Performance Indikator)
for instance, facebook based on register user, because it's tend to be a social network
Air BnB, based on net booking
how many hotels had been booked everyday
and we could apply this, in every bussines sector
for example, e-commere. How many sale eachday
next, I want to introduce the required cycle
decision start from data
we need to define which data that could be used on our hyphoteses
how to drive our company to increase, according to KPI
based on data, we could build hyphoteses but sometimes, it's not enough
there is need another insight
that we could from research (Customer interview, survey, or market research, etc)
after we collect the data, insight, and hyphoteses
we could get optional decision
what we can do in the future ?
and then take action
because decision whithout take any action won't be anything
next, we evaluate our action based on data
to check whether it's leads to better direction
so, we could decide which one that can be maintained or eliminated
next…
with the existence of developing technology, we can join a lot of datas
make it insightful
firstly, strategic team should aware about business trend whether it's going well or not
We could check from reporting dashboard
To create an actionable and insightful reporting dashboard for stakeholder is not easy
Then.. There is analysis
Analysis can be divided into historical analysis, exploratory analysis, and forecasting 96 00:10:03 --> 00:10:05 We discuss about reporting dashboard
Firstly, before we create a dashboard we need to determine who going to receive the dashboard
Because no every stakeholder get the same dashboard
For example, operational team. The dashboard would portray customer experience, complaint (how many complaint), how long the lead time
The goal is to build the bonding with the customer
So the customer will be loved to use our product
Because, company is built for customer
Secondly, when the dashboard will be sent
For example, operation team
if they need an alert, they can get it in a real time so they can make a quick handling
it going to be different with management level might be they don't need a real time, but they want to overlook weekly pattern or monthly
so every stakeholder will be different
for instance ad-hoc, they don't need regular basic
example for marketing, they need to build dashboard or data analyst if they want to checkthe impact of the promo to new user
the las one, what is the key matrix
again emphasized that support KPI it is not necessarily have to shared or used data
or we need to know how to process data that can be used
we need to use a proper matrix
What does mean with a proper matrix?
(Lyn) analytics book, there are four things for matrix refers to parameters
First, comparative
For example, last week's total sales were 1000
That's it? Without any comparison? It will not mean anything
We could add the data
Last week's sales were increased 10% compare to the previous week
This is could shows that performance is increased
If its decreased, we need to deep dive the cause
Understanable... it is useless to build a matrix to someone that does not understand
The selection of the legends must be precise
So.. it will be understanable
Definitons must be equalized
For Instance, definiton 'Turn' (please check the term) is not used by some people
Whether it is still in the platform or they do not booked it
Next... Ratio or rate.. Matrix is better in ratio instead
For Example
We want do a benchmarking between product A, B, and C
Eventhough, all the products are rising
But, we have to examine the market share for each product
For example, product A is rising until 80%
We can decide to focus on selling product A
Without putting more effort into a smaller rising product
Let's talk Behave Changing
It is tend to how chainable your matrix
As a benchmark for making a decision or hyphoteses
For Instance, this is not an actual data
For Instance, Total Monthly Sales... Which one do you prefer?
Which one do you prefer? Right or left side?
Are these data simillar?... Yes, They are
Which do you prefer?
The data on my left or right side?
Hmmm... everyone still choose different side
Actually, there is no right or wrong answer
It depends on your Goal
Lets says... we just launching a premium item
And we want this item become an idol
Desired by people
And turned people who used regular item into the premium one
We can see the increase in both charts
But the chart on the right, can shows its proportions
It shows the proportion for premium items are increased
Which is good for us
And we can see, how much its increased
From point 1 to 2, its rising quite high
So we can explore, what is the cause from point 2 to 3
Maybe.. we can repeat doing a campaign
Which can rise the chart from point 2 to 3
Would the real matrix could be actionable?
First.. we have total data of sign up
We can detect, how much its increase per day
But, we can not do any action
Sign up could shows active user
Because from the awareness, we want active user more active
This is more make sense
It is going to be better if we add the ratio
Below.. the real matrix for strategic decision
Example, from MAN 1 to MAN 5 (please check the term) the percentage are decreased
Based on that we can decide, which one can be fixed
To increase active user
Why active user always decrease?
We can make cohort
Cohort has a lot of varians
In here, i show you cohort for retention rate
What it does mean?
We can see the growth of new member in january
It shows how many exist-members are using the platform in the second month
In the third month, it shows how many (in percentage) exist-members are using the platform
We could track the member who sign in february and march
Sometimes, if there is a different initiative the cohort will differ as well
Cohort can make an action
It shows the number from point 1 to 2 is drop
We can explore
This report rises hyphotheses and question
This is what data analyst should do
Always curious
Next.. Funnel
Funnel is used for mobile analytic
It could detect when people stop doing 'klik' or order
It is also shows a drop point from one action to another
From point 1 to 2, we can see... it shows quite a lot
From point 2 to 3, view sign up page to sign up
It is identify less than half of members do sign up
How we encourage people to sign up?
After we identify the data, we could know the number
We should know, how many our user
MAU (Mainly Active User)
We could know, which product are they used
So we could feel the empathy as a customer
Second, Listen and Explore from external data
Because Internal data is considered not enough
We could identify from customer's feedback
Its origin from the suctomer
Or we could used social media
Ask question if there is any drop performance
We establish a fishbone
Based on the cause of drop performance, whether its cities or products
Collect data and analysis
Brainstorm
For example, business analysis in GOJEK
The business analysis is under business unit
GoSend has its own business analyst
I work in GoPay business unit
In here, we work in-line wit BI
BI supported in data clearing and etc
We work in-line with growth, marketing, etc
In order to do brainstorming
Next, exploratory analysis
Firstly, after all the dashboards we could use intuition
Because it is important
And we could use a lot of methods
For example, we reckon there are loyal customers and low users
We could establish a cluster to analyze how many times low user should using the platform to be called high user
Then explore 'why'
Because the data only could answer 'what' and maybe 'how many/much'
So we could provide the appropriate to the customer
The last, combine quantitative ang qualitative analysis
This is why we always do brainstrom and research
Because a lot of insight from others
Which is has ot described in our report The last about strategic decision
Actually, if we talk about strategic decision
There is a senior level as a decision maker
In here, we establish from data
In order to establish the data, we should know the strategic based on KPI
And we have to build the framework
Because when we facing a lot of datas, we become more cautious and not focus
So we need to be strict and focus
Which part that we going to build
And then, breakdown the hyphotheses
What are the cause that make GoRide drop? Whether is the areas, products, or wheater
Validate data experiment
This experiment could us AB testing, Multivariate Analysis, and etc
To decide the variables
Decide the actions
Strategic level is more toward long term strategy
So.. if they want to establish a campaign, pricing scheme, etc
They going to overlook cake and customer accuisation class (please check the term)
It is shows how much the cost that have been spent
In the begining we offered a lot of discount, promo
We could define how many new users that we can gain and join with us
But.. do not forget to overlook the retention rate
We need to check, which user that has a good retention rate
Based on that, we could see LTV (Life Time Value) Customer
After the customer has been engaged with us
We could decide what we can offer
And how long it can be used by the customer and how much revenue is generated
Last... evaluate result
We must to do monitoring for every single decision
We could build another report or analysis
For more infomation >> Tina Fey Addresses 30 Rock Reboot Rumors - Duration: 4:40. 
For more infomation >> Tina Fey Debuts Her Daughters' iMovie Film "Butt Show" - Duration: 5:42. 
For more infomation >> Industry responding after fatal Southwest flight - Duration: 1:47. 
For more infomation >> Verschleppt und gefoltert - Dschaisch al-Islams "Gefängnis der Reue" in Duma enthüllt - Duration: 1:35. 
For more infomation >> Late April snow disappoints many Granite Staters - Duration: 1:28. 
For more infomation >> Observant neighbor films mean break into home - Duration: 1:22. 
For more infomation >> Not great, but a decent weekend forecast - Duration: 3:01. 
For more infomation >> Santa Cruz Sheriff: Boy, 15, was killer behind Soquel Drive homicide - Duration: 0:30. 
For more infomation >> Freeze Warning Overnight - Duration: 3:24. 
For more infomation >> With marijuana legal, what will 4-20 in Santa Cruz be like? - Duration: 2:46. 
For more infomation >> Man accused of beating woman in 2008 pleads no contest - Duration: 1:51. 
For more infomation >> நிர்மலா தேவி பேசிய மோசமான வார்த்தைகள். - Duration: 1:49. 
Không có nhận xét nào:
Đăng nhận xét