Data Analyst on How to Turn Business Metrics to Insights

0:00so a lot of aspiring data analysts are

0:01missing a magic ingredient when it comes

0:03to standing out in the job market and

0:05that is a solid handle on how to use

0:07business metrics and translate them into

0:09real insights and recommendations that

0:11resonate with a stakeholder if you're

0:13new here I'm Christine I'm a former data

0:15director and hiring manager and I'm also

0:17founder of the analytics accelerator

0:19program in which over 70% of students in

0:21my first cohort landed jobs in data

0:24within just 6 months of the program

0:25ending if you also want to go from

0:27aspiring to S out make sure to hit the

0:29Subscribe button and also the

0:30notification Bell so you know when a new

0:32video is out so in this video I'm going

0:33to give you a crash course on why

0:35metrics are the backbone of being a

0:37strong data analyst and then we're going

0:39to take a bird's eye view and look at

0:40some of the metrics that you should know

0:41across a variety of different Industries

0:43and actually walk through a framework

0:45for how to translate a metric to a

0:47recommendation and then I want to show

0:49you what this looks like in action with

0:50a concrete example I would love to

0:52actually meet you guys live so I'm going

0:54to be hosting an open Q&A this Saturday

0:56where I'm happy to talk more about

0:58business metrics and some my insights

1:00about the job hunt that I've seen work

1:01really well for my students so we all

1:03know that a data analyst turns data to

1:05insights and recommendations and if we

1:08look one level deeper we'll see that

1:09insights are really just about

1:11understanding metrics and their root

1:13causes and effective recommendations is

1:15built on top of effective communication

1:18in this case metrics are pretty much the

1:20what of what we're trying to understand

1:22on the day-to-day and the root causes

1:24explain the why behind their movements

1:26and their fluctuations if we think about

1:28the day-to-day question that we're going

1:29to be getting from from stakeholders

1:31those are questions like from a

1:32marketing manager why did the conversion

1:34rate on the US free trial page go down

1:36or maybe you're working with an fpna

1:38analyst and they're asking what were the

1:39main drivers behind the revenue Spike

1:42last month or maybe you're working with

1:43a product manager who's asking what is

1:45the adoption rate of the new webinar

1:47tool and how does that compare against

1:49the old webinar tool at first all of

1:51these questions can look pretty

1:52overwhelming but if you take a step back

1:54you'll see that they're pretty much all

1:56asking about a metric and that metric's

1:58movement and at the end of the day the

2:00big question is always what should we

2:02actually do about it impactful data

2:04analyst help people understand the so

2:06what behind why a metric is moving up or

2:08down and so the more you can actually

2:10show this ability in your projects and

2:12in your interview responses the more

2:14immediate value they'll see that you can

2:15bring to the team and the moreal stand

2:17out so I want you to actually start

2:19familiarizing yourself with the most

2:21popular metrics that data analysts work

2:23with and I'll just start you off by

2:24sharing some of the metrics that I've

2:26worked with throughout my own career

2:27when I was working at just works as a

2:28data analyst I part with the marketing

2:30team and they cared about metrics like

2:32traffic Impressions click-through rate

2:34and customer acquisition cost then I

2:36worked at Oscar Health which is a

2:38healthcare and tech company and they c r

2:40metrics like Revenue per member claim

2:42amounts claim counts and medical loss

2:45ratio when I was working at Vimeo I

2:46worked with a finance team where they

2:48cared about metrics like recurring

2:49Revenue subscribers and bookings and a

2:52product team which cares about numbers

2:53like active users or net promoter score

2:56and customer satisfaction score now a

2:58lot of aspiring analysts ask me well

3:00what if I don't want to pigeon hole

3:01myself into one specific industry or

3:04what if I'm not really sure what

3:05industry I actually want to work in how

3:07do I gear my projects towards something

3:09that will resonate with a hiring manager

3:11in that case if you look back at the

3:12stakeholder question that I showed you

3:14earlier you can see that it pretty much

3:15revolves around what drives a metric up

3:18or down across various dimensions and

3:20what should we do about it which means

3:22that if you can understand the

3:23analytical techniques to actually

3:25translate metrics to insights you can

3:27pretty much abstract away what that

3:29speciic specific metric or where that

3:30Dimension is yes domain knowledge is

3:33super important and it can really make

3:34you stand out but for early career data

3:36analysts it's most important to learn to

3:39apply the Frameworks that turn these

3:41metrics into insights because then you

3:43can use that to stand out across a

3:45variety of different Industries so I'm

3:47going to show you a really simple

3:48four-pillar framework that I use to

3:50understanding metrics and first you want

3:52to think about what are the questions

3:54that we're actually trying to answer in

3:56this case it's what metrics do we care

3:58about what are their values and their

4:00Trends what drives their movement and

4:02what should we do about it and each of

4:04these questions can be broken down into

4:06smaller and smaller steps the next level

4:08down are doing things like defining the

4:10Northstar metrics and dimensions and if

4:12you don't know what I mean by Northstar

4:13metrics make sure to watch my data lingo

4:15video then the next step is to report

4:18visualize and slice and really start to

4:19explore and understand what some of

4:21those Trends are so that you can start

4:22to uncover the drivers and relationships

4:25that actually impact that metric moving

4:27up or down lastly that's when you start

4:29to formulate your recommendations based

4:31on what you saw and communicate that to

4:33stakeholders in an understandable way we

4:35can also break this down more into

4:36smaller steps where the first step is

4:38pretty much defining and calculating

4:40each key metric now I say Define because

4:43you'd actually be surprised the number

4:44of times where at work we think we're

4:46talking about the same metric but it's

4:48actually calculated differently across

4:49different teams so it's really important

4:51to document what those metrics are and

4:53how they're calculated then identifying

4:55the associated key Dimensions so finding

4:58the qualitative segment that you have

5:00available in your data set that you can

5:02slice that metric by just to understand

5:04how that metric looks across the entire

5:05population then in reporting visualizing

5:08and slicing we can almost always slice

5:10by a universal dimension called time and

5:13that's just the start of doing something

5:15like seasonality analysis then when we

5:17slice by those other dimensions and the

5:18other qualitative values that's where we

5:20start to get into segmentation and

5:22through this process once we're looking

5:24at those values and what we've built we

5:26can start to actually translate those

5:28Trends to observations and report those

5:30values and findings normally as a data

5:33analyst we be working on a team where

5:34there's a larger context about how these

5:36findings fit into the bigger picture

5:38where you would work together to start

5:39to craft a digestible narrative with

5:41other people who have contexted about

5:43how this fits into the other parts of

5:45the business so let's take a quick look

5:47at all this in action let's say I'm

5:48working as a data analyst at a

5:50Healthcare company and a marketing

5:51manager ask me what are the best

5:53performing campaign types based on

5:55signups and what should we actually do

5:56about this in this case my data set

5:58looks something like this where the

5:59table green is the campaign ID then I

6:01have campaign category campaign type

6:03which is just a subset of the category

6:06the cost the platform The Impressions

6:08clicks Days Run and the number of

6:10signups that trickle down from that

6:12campaign because the stakeholder has

6:14defined best based on signup metrics I'm

6:17going to identify myor star metrics as

6:19sign up rate cost per sign up and sign

6:21up count in terms of the dimensions I

6:23have campaign category campaign type

6:25platform and in this case time got cut

6:27off but we also have a Time variable so

6:29one of the first things I would do is

6:30actually quite simple I would just

6:32calculate those metrics and slice it by

6:34one of the first key Dimensions which is

6:36the campaign category and in this case

6:38if I were to sort everything by signup

6:40count I can see that healthy living

6:42actually has the most number of signup

6:44counts but health for all which comes in

6:46second place in terms of the absolute

6:47value has a much higher signup rate and

6:50therefore a much lower cost per sign up

6:53interestingly the worst performing

6:55campaigns they are not only the worst in

6:57terms of the signup count but across the

6:59board with the other sign up metrics as

7:00well so that would definitely be

7:02something to look more into the next

7:03thing I would do is actually visualize

7:05this over time and so if I build a line

7:07graph and I can look at what these

7:08different campaign categories look like

7:10over time with a legend of course I

7:12would be able to see a little bit more

7:13but I can definitely see that there's a

7:15huge spike in the beginning of 2020

7:17probably coinciding with covid because

7:19this is a Healthcare company and then at

7:21the end of 2021 and 2022 when every

7:23other campaign is performing worse the

7:26campaign that's in Orange is actually

7:27still on the rise until it goes way down

7:30at the end of 2022 so definitely

7:32something to look into there as well

7:33right now when I'm just getting started

7:35I'm just prioritizing Clarity over

7:37complexity I'm not necess doing

7:39something really fancy at this point

7:41because I'm really just trying to

7:42understand what those overall Trends are

7:44in the life of this metric and even with

7:46just this simple table and this graph I

7:49can already start to surface some

7:50insights like this one the hash healthy

7:53living social media campaigns had the

7:54highest number of signups but health for

7:56all campaigns have significantly higher

7:58sign up rate with half the cost of

8:00acquisition and if I were to bring this

8:02to a stakeholder we probably have a

8:03discussion about what this means in

8:05terms of the budget I could start with

8:06something like consider reallocating the

8:08budget from the three worst performing

8:10campaign types which perform much worse

8:13across all three northst metrics to

8:15increase the budget for health for all

8:17campaigns which have the highest Roi

8:19early on in your career as a data

8:20analyst you probably won't be

8:22responsible for giving the end all Beall

8:24recommendation to another team instead

8:26You' be working on a team to bring

8:27together the context of what's actually

8:29feasible and marry that with what you're

8:31observing in the numbers to figure out

8:33what the next steps are so if we look at

8:35what this actually looks like in a

8:37project in GitHub you could put this in

8:39an Insight summary where in this case

8:40this student identified his Northstar

8:42metrics as signup rate cost per sign up

8:45and click-through rate and then he also

8:46went through each of those metrics and

8:48started observing what are the top line

8:50insights that someone would want to know

8:51about so in this case he talked a little

8:53bit more about the signup rate and then

8:55later on in his project went way more

8:56into detail about the visualizations the

8:58recommendations and the technical

9:00process so just give you a little bit of

9:02motivation for how this can help seal

9:03the deal in an interview process this is

9:05a note that one of my students got after

9:07he did an amazing job at an interview

9:09and they said you did such an amazing

9:10job preparing for the day and organizing

9:12your thoughts and examples in an easy to

9:14follow structure thanks for sharing

9:16links to your project work these are

9:18wonderful examples of distilling data

9:20into meaningful actionable insights and

9:23that is what people really want so we

9:25pretty much just skim the surface when

9:26it comes to understanding business

9:28metrics but I really hope that gives you

9:29a better understanding about how this

9:31all fits in to the day-to-day job of a

9:34data analyst I'd love to interact with

9:35you guys live beyond just YouTube so

9:37check out the link in the description

9:39below for a free Workshop I'm hosting on

9:41Saturday to answer some questions about

9:43all of this and also talk more about how

9:45you can get direct mentorship and

9:47Community along your journey if you also

9:49want to go from aspiring to S out don't

9:51forget to subscribe and I'll see you

9:52guys soon