Full transcript
Introduction
0:00As a data analyst or someone
0:01transitioning into data, have you ever
0:02had this moment where you're in an
0:04interview or in a stakeholder meeting
0:06and someone says, "Yeah, the data looks
0:08right, but what does this actually mean
0:10for our business?" And you have a small
0:12bit of panic because you know that your
0:14dashboard is built correctly, you know
0:15your SQL is clean, and your charts are
0:17accurate, but the story behind the
0:19numbers feels fuzzy. I see this in
0:21analysts that I work with all the time.
0:23They can build a dashboard in their
0:24sleep, but the moment someone asks, "So,
0:26what's the so what here?" then they feel
0:28like they can't answer this question
0:30like a real data analyst. If this
0:31applies to you, you have what I call a
0:33business language gap, which means you
0:35can overcome this by starting to
0:36understand the thinking [music] behind
0:38business metrics. So, today we're going
0:40to cover the top 10 metrics that every
0:42data analyst should know, grouped into
0:44four key families that is going to help
0:45you form a mental model of all business
0:48metrics, so they can go from sounding
0:50like you're just a dashboard to actually
0:52being able to tell the so what and the
0:53story behind the numbers. And stick with
0:55me cuz at the end, I'm going to give you
0:57a pro tip about how to sound like an
0:59experienced analyst before you step into
1:01an interview. If you're new here, I'm
1:03Christine. I spent a decade in data as
1:05an analyst and then became a hire
1:07manager and a data director, and now I
1:09help working professionals and data
1:11analysts stand out in the job market and
1:12on the actual job through my mentorship
1:14program. If that sounds helpful to you,
1:16make sure to check that out down below
1:17and let's dive in. So, the first family
1:19of metrics is money coming in, and the
1:21first metric here is revenue, which is
1:23total dollars in. Now, we all know that,
1:25but what smart analysts understand is
Metric 1
1:27that while revenue is a North Star
1:28metric that everyone in the company
1:30cares about, the story behind revenue is
1:33much more telling than the top-line
1:35revenue number itself. That's because
1:37revenue on its own doesn't actually tell
1:39you at all about the health of a
1:41business. Revenue could be going up
1:43while costs are going up even faster,
1:45which, fun fact, is not actually a
1:47success story. So, when someone says
1:48something like revenue has risen by 15%
1:51this quarter, a smart analyst knows that
1:53the next question isn't just great,
1:54"What's next?" It's, "Where is that
1:56growth actually coming from and is that
1:58growth sustainable?" A quick note on how
2:00this number shows up across industries.
2:01[music] In SaaS or anything selling some
2:04kind of subscription or software, you
2:06might hear something like ARR or MRR,
2:08which stands for annual recurring
2:10revenue or monthly recurring revenue.
2:12For any uh company that is selling a
2:14physical product, you might hear
2:15something like total sales or just
2:17revenue. And for a consulting company,
2:19for example, you might hear something
2:20like total project billings. It's
2:22slightly different terminology across
2:24industries, but they all mean the same
2:26concept. So, revenue is always a good
2:28number to know, but the real health
2:30indicators behind it are the customers
2:32and also how much each customer is
2:34worth, which is what we're going to talk
2:35about next. So, metric number two is
2:36ARPU or AOV. So, if revenue is the total
Metric 2
2:39dollars coming in, then ARPU or AOV is
2:41the value of each customer or
2:43transaction. So, in a SaaS business or
2:46anything that's selling some kind of
2:47subscription, you might hear something
2:48like ARPU, which is average revenue per
2:50user. In something selling products like
2:52e-commerce, you might hear something
2:54like AOV, so average order value. And in
2:57anything that's B2B, so a company that's
2:59actually selling to other businesses,
3:00let's say a consulting firm, you might
3:02hear something like ACV, which is
3:04average contract value. Again, different
3:06acronyms, but same idea. How much is
3:08each customer or transaction bringing
3:10in? Now, here's why this detail matters.
3:12Let's say Apple increased their revenue
3:14by 20% in one quarter. Sounds great. The
3:17finance team sounds really happy. They
3:19send an enthusiastic Slack message. But
3:21then, if you look underneath the hood,
3:22you see that their average order value
3:24has actually dropped by 25% in that same
3:27time frame. What does this actually
3:29mean? It means that they started selling
3:31more of their cheaper products, so let's
3:33say AirPods, products that are on the
3:34lower end of their price list compared
3:37to their more premium products like
3:38MacBooks. So, even though revenue went
3:41up, the product mix actually got weaker,
3:43which isn't a good sign in the long run.
3:45Now, this is one of the first things
3:46that we look at when we're hiring
3:47analysts, which is not just, "Can he
3:49report on this top-line number?" but
3:50also, "Can you understand the main
3:52drivers behind what contributed to this
3:54top-line number and why it changed?" So,
3:56average order value and ARPU is usually
3:58one of the first places we start when
4:00we're thinking about how revenue has
4:01gone up or down. By the way, if you want
4:03to shortcut your understanding of
4:04business metrics and be able to speak
4:06about them like an experienced analyst,
4:08then I highly recommend you grab my
4:09business metrics cheat sheet down below.
4:11It's probably going to save you like 20
4:13hours of Googling cuz it covers the top
4:15metrics that every analyst should know
4:16across specific industries as well as
4:18how to speak about them so you sound
4:20like a knowledgeable analyst from day
4:21one. So, make sure to grab that down
4:23below. So, now we've covered money
4:24coming in. Time for the less glamorous
4:25question, which is, "How much of that
4:27money actually goes out and how much of
4:29this money do we actually get to keep?"
4:30So, metric number three is gross margin.
Metric 3
4:32Gross margin is what you keep after
4:34subtracting out the cost of delivering
4:36your product or service. So, for
4:38example, let's take Netflix. Netflix
4:40charges each user $15 per month, and
4:43let's say that the cost of actually
4:44hosting and delivering and streaming the
4:46videos is like $8 per person per month.
4:49So, they get to keep seven out of 15 of
4:52every dollar that they charge, which is
4:54about a 46% gross margin. Now, here's
4:56what I find pretty interesting is that
4:59even though companies have a the same
5:00top-line revenue number, let's say two
5:02companies both post $5 in revenue in one
5:05quarter, but they can be in totally
5:07different financial health because a
5:09SaaS company that has 80% gross margin
5:11is doing much better than a sports
5:14retailer that only gets to keep 25% of
5:16their revenue. So, when I was hiring
5:17candidates, one question we used to ask
5:19was, "If revenue is going up, but gross
5:21margin is shrinking, then what would you
5:23look into next?" And the analysts who
5:25stood out always understood that gross
5:27margin shrinking when revenue going up
5:29usually means that something is
5:31happening with the expenses, the costs,
5:33or discounting or promotions. And they
5:35knew that they would have to loop in the
5:37finance, product, or marketing team to
5:39actually investigate it further. Now, in
5:40terms of how this applies across
5:42industries, so SaaS margins are usually
5:45pretty high because the cost of
5:47delivering the service is usually just
5:48around the actual technology. Whereas
5:51for a retailer or someone actually
5:53selling a physical product usually have
5:54smaller margins because they have a lot
5:57of costs around inventory, delivery,
5:59operational costs, and also returns also
6:01weigh into that. And for consulting, the
6:03main cost is the actual people that they
6:05are employing to work on these projects.
6:07So, across industries, we all have our
6:09different benchmarks again of gross
6:11margin, but the main question is the
6:13same. How much does it actually cost to
6:15deliver one unit of value and how much
6:17are we keeping after that? Now, one of
6:19my students actually used this
6:20understanding of business metrics to not
6:21only land a new job, but pivot into an
6:24analytics role. And I want to share his
6:25story with you just as a bit of
6:26motivation. So, he had been laid off
6:28from Microsoft about a year ago, and
6:30when he was interviewing for a labor
6:31analyst role at Palms Casino Resort, he
6:34actually created a forecasting model
6:36that demonstrated his understanding of
6:38their top metrics. So, it was labor
6:40hours, uh labor costs, and other
6:42resources, and he showed that model to
6:45the hiring manager in an interview. And
6:47he used to say the manager was super
6:48impressed by his domain knowledge even
6:50though he's coming from another
6:51industry. And when he got that offer, he
6:53actually told me that he'd been laid off
6:54from Microsoft a year ago, and so
6:56pivoting into new industry and also
6:58getting an analyst job feels like a
6:59double win. So, that just goes to show
7:01you how much understanding business
7:03metrics can be a game-changer when it
7:05comes to job hunting and interviews. So,
7:07metric number four is CAC or customer
7:09acquisition cost. It's basically just
7:11how much it costs to bring in a new
Metric 4
7:12customer. So, the formula here is total
7:15spend used to acquire new customer
7:16divided by the number of new customers
7:18you got in that period of time. Now, it
7:20sounds kind of straightforward, but
7:21realistically on the job, finance,
7:23marketing, and leadership all have to
7:25agree on what is actually captured in
7:27this numerator. So, does spend also
7:29count team salaries, tools, events, or
7:32does it only count marketing spend? And
7:34so, usually these definitional changes
7:35can make a really big difference in how
7:38this metric actually looks. So, on the
7:39job, it's really important to understand
7:41exactly how each of these metrics are
7:43actually defined because if you change
7:45the definition, it's really easy to make
7:46this metric go up or down. So, going by
7:48our Netflix example, let's say they have
7:50a $90 CAC. So, they spent about $90 on
7:53average acquiring a new customer. Now,
7:55again, this is just an example, so I'm
7:56just making up the numbers here. But is
7:58this number good or bad? To be honest,
8:00on its own, you cannot really understand
8:02whether a company's CAC is good or bad.
8:04It almost always has to be looked at in
8:06comparison to another metric, which
8:08we're going to talk about next. So,
8:09family number three is customer
8:10behavior, which has a metric that is
8:12going to be related to CAC, which we're
8:14going to talk about in a little bit.
8:15Before we dive in here, it's important
8:16to understand the different business
8:18models when it comes to who companies
8:20actually sell to. So, there is B2C.
8:22That's when people are selling to actual
8:24individuals or customers, so think of
8:26companies like Sephora or Lululemon.
8:28They're selling to an individual person.
8:29Then there's B2B, so companies that sell
8:31to other businesses. So, any kind of
8:33SaaS company that has like an enterprise
8:35solution, for example, Vimeo Enterprise.
8:37They sell to other companies, and so
8:39they're counted as B2B. Then there are
8:41also companies that are technically
8:42B2B2C.
8:44Yes, it's a real term. I did not make
8:45that up. For example, I have a student
8:46right now who works at Penguin Random
8:48House as an analyst. And uh Penguin
8:50Random House creates and publishes books
8:53that obviously go to individuals, but
8:55they also mostly sell to large retailers
8:57like Amazon and Barnes & Noble's. So,
8:59that's why I would consider them
9:00somewhere in the middle, so B2B2C. So,
9:03whenever you're interviewing for a
9:04company or onboarding onto a new job,
9:06it's really important to understand who
9:08they actually sell to and what business
9:10model they're a part of. That's really
9:11good background context for the metrics
9:12we're going to talk about next. So,
9:14metric number five is lifetime value,
9:15which is the total expected value or
9:18revenue that a customer is expected to
9:19bring in over their entire relationship
Metric 5
9:22with a business. So, back to Netflix.
9:24Let's say that on average a customer who
9:26pays $15 per month is expected to stay
9:29with Netflix for 3 years or 36 months.
9:32Now, I'm sure there are many, many
9:33outliers on the higher end, you know who
9:34you are, but on that case, the
9:36customer's lifetime value is $540.
9:39Now, compare this to ARPU, which we
9:41discussed earlier. The ARPU is estimated
9:43to be $15. So, ARPU is the money that
9:46you're actually coming in with today,
9:47whereas lifetime value is the money
9:49you're expected to get from that
9:50customer over the entire time that they
9:52stay with you. As you can see, the
9:54customer acquisition cost or the CAC
9:56that we discussed earlier only makes
9:57sense when you actually compare it to
9:59how much you're expected to get from
10:00this customer over the time that they're
10:02with you. So, in our Netflix example,
10:04when we had a $90 CAC and the LTV is
10:07$540, we want to calculate what we call
10:10the LTV to CAC ratio. In this Netflix
10:12example, that LTV to CAC ratio is about
10:14six. And a healthy industry benchmark is
10:17around three to one. So, in this
10:19example, Netflix would be doing really,
10:21really well. So, if you're an analyst
10:23working with any kind of product,
10:24marketing, or like product operations
10:26team, you probably will hear about this
10:28ratio, LTV to CAC, because it is one of
10:30the strongest signals of whether or not
10:32a company's growth engine is actually
10:34working. So, magic number six is the
10:35number of customers. Now, this sounds
10:37like it's probably something that's
10:38really easy to calculate, but let me
Metric 6
10:40tell you it is not. So, one of the first
10:42questions is usually definitional, which
10:44is just who actually counts as a
10:46customer. If we are working with a
10:48subscription business, then what counts
10:50as being actually active and on the
10:51platform? Do we count free trial users?
10:54Do we count anyone who signed a
10:55contract? If they left the platform and
10:56they came back, then at what point do we
10:58count them again? So, as you can see,
11:00there are many different ways to
11:01actually count this value of number of
11:03customers. And at work, many teams are
11:05often using very different definitions
11:07of the number of customers without
11:08actually realizing it. Now, as an
11:10example, let's take subscription
11:11businesses. They will have a few
11:13different versions of this customers
11:15number that they're tracking. So, we
11:16will track things like newly acquired
11:18customers. We'll track things like
11:20churned customers. And from the two
11:21numbers, we will track net customers.
11:23So, in this case, we'd actually call
11:24them subscribers. So, newly acquired
11:26subscribers, churned subscribers, and
11:27net new subs. As you can see, on its
11:30own, the top line number of number of
11:31customers doesn't really tell you the
11:33full picture. You need to look at things
11:35like LTV, CAC, and churn to start to
11:37understand the story behind the numbers.
11:39Magic number seven is NPS or net
11:41promoter score. And don't make the
11:43rookie mistake of saying NPS score,
Metric 7
11:45which is net promoter score score. So,
11:47all of the metrics that we've been
11:48covering so far give us a picture of a
11:50company's financial health, but none of
11:52them answer this question, which is of
11:54all the customers that you have, how do
11:56they actually feel about you? So, net
11:58promoter score is calculated by asking
12:00people, "Well, how likely are you to
12:02recommend us to a friend?" And then we
12:03take the percent of people who answered
12:05highly and subtract the percent of
12:07people who answered with a really low
12:08score. And then we convert that to a
12:10scale of -100 to 100. Netflix's NPS sits
12:13at around 50, let's [music] say, which
12:15is pretty good. And that means that for
12:17every one of their customers who then
12:19brings in another friend, that's a
12:20customer that they just acquired for
12:22free. So, high NPS actually brings CAC
12:25down. lower your NPS is, the more it's a
12:27leading indicator that revenue might
12:29drop and churn might increase because
12:31usually, if satisfaction goes down, then
12:33the dollars you earn and the number of
12:34people who stay with you also goes down.
12:36So, that's how these metrics are related
12:37to each other. In practice, on the job,
12:39usually we also wouldn't look at just
12:41one high-level NPS. We would also break
12:43NPS down by some key customer
12:45dimensions. So, we might slice it by
12:47region. We might slice it by age or any
12:49other demographic information that we
12:50understand, so that we can start to
12:52relate this to a customer experience,
12:54marketing, or product team. So, family
12:56four is funnel health, which in my
12:57opinion are some of the more fun and
12:59interesting metrics to work with. It
13:01tells you about the customer journey
13:03from going from a stranger to becoming a
13:05loyal paying customer. So, metric eight
13:06here is the concept of the conversion
13:08rate. So, at any step in the customer
Metric 8
13:10journey, there are milestones that
13:12happen to move from becoming a stranger
13:15who wasn't aware of your company in the
13:16first place to becoming a paying
13:18customer. For example, for Netflix,
13:20they're definitely tracking the number
13:21of people who land on their homepage and
13:23actually sign up for a paid account. And
13:26then for an e-com company, let's say
13:28Sephora, they're tracking the number of
13:29people who land on a targeted ad, then
13:32they click on that ad, they go to the
13:33website, then they track the percent of
13:35people who on that website actually add
13:37something to their cart, and then they
13:38track what percent of those people
13:40actually check out their cart and
13:41fulfill their order. And a consulting
13:43company, for example, is going to track
13:45the number of initial conversations they
13:47have that actually turn into paid
13:48contracts. So again, it's all about
13:50tracking the percent of people that move
13:52through each step of starting from being
13:54a stranger to becoming a paid customer.
13:56And every business has this concept of
13:58the funnel. So, this is where we go from
14:00that top line question of what's
14:02happening in revenue to actually being
14:04able to dive into what's driving those
14:06changes in revenue, what's a bottleneck
14:08there, and being able to answer that
14:09question, what should we actually do
14:10about it? Here is a tip for the next
14:12time you have an interview, is to go to
14:15a company's website and actually map out
14:17their customer journey as if you were
14:19going from a stranger to a paid
14:21customer. So, what is their funnel? What
14:22are all the different milestones in that
14:24journey to go from being an unaware
14:26customer to actually checking out or
14:28paying for their services? And what are
14:30the metrics that you would actually
14:31track to understand the performance at
14:33every single step? Now, most people do
14:35not do this. So, if you understand a
14:37company's funnel before you step into
14:39that conversation and it shows, you're
14:41already going to sound like someone
14:42who's been working in that job for a few
14:44months. So, metric nine and 10 are
14:45actually two sides of the same coin, and
14:47they are retention and churn. So, while
14:49conversion tells you how effective you
Metric 9 & 10
14:50are at bringing people into the
14:51business, retention and churn tell you
14:53how effective you are at keeping them.
14:55So, while retention is the percent of
14:56people who stay with you, churn is the
14:58flip side of that, it's the percent of
14:59people who leave. So, let's take Netflix
15:01again as an example. They spend millions
15:03of dollars every year investing in
15:05creating new content, creating new
15:06shows, because they know what it does
15:07for customer retention. If you are
15:10signing on to Netflix at 11:00 p.m. to
15:12watch the last three episodes of your
15:14favorite show and you're getting hooked,
15:15you're obviously not opening the app to
15:17then cancel and churn. And Netflix knows
15:19this, so they invest accordingly. So, in
15:21an interview or a stakeholder
15:22conversation, especially if you're
15:24working with anyone close to the product
15:25or operations side, it can be really
15:27insightful to ask questions about
15:29retention and churn, because it will
15:30help you understand who their customers
15:32are. So, a question like, uh what are
15:34trends in retention recently and how
15:36does that determine some of the
15:37strategic priorities at the company? Or
15:39something like, what are trends in
15:41churn? And how has the team used data to
15:43try to drive decisions and
15:45recommendations to help address it? So,
15:47these are just some examples of really
15:48good questions that you can ask at the
15:50end of interviews or in coffee chats to
15:52signal your understanding of business
15:54metrics, but also show that you really
15:56care about the customer base at the
15:58company. So, retention and churn are
15:59definitely metrics that anyone working
16:01close to a product or customer
16:03experience team will need to know. So,
16:04those are top 10 metrics that I
16:05recommend starting with. As you can see,
16:07it's not about memorizing formulas,
16:09because the metrics are going to differ
16:11based on what kind of company and role
16:12you're actually working in. It's much
16:13more about a way of thinking. So, um you
16:16know, how's money coming in? How's money
16:17going out? How much of it are we
16:19keeping? Who are our customers? And how
16:21quickly do they become a customer and
16:23how much do they actually stay? The more
16:25fluidly you can move between these kinds
16:27of questions, the more naturally you're
16:29going to be able to speak the language
16:30of business, the more you're going to be
16:31able to actually tell the story behind
16:33the numbers instead of sounding like a
16:34dashboard. That's what companies are
16:36really looking at today. As I mentioned,
16:37I put together a free business metrics
16:39cheat sheet for you guys, so that you
16:40can learn these metrics and much more
16:42depending on whatever industry you are
16:43interested in applying to. So, if you
16:44want to save a lot of time Googling,
16:46make sure to download that down below.
16:47And if you're interested in taking this
16:49a step further and really understanding
16:51business metrics in a structured way, so
16:53that you can stand out in the job market
16:55and on the job, then check out my
16:56mentorship program down below. I'm going
16:57to leave a link at the bottom. And
16:59that's it for this video. I will see you
17:00guys in the next one.