Adora Cheung - How to Set KPIs and Goals

00:00

all right so i am going to be talking

00:02

about setting your kpis and goals

00:05

for early stage startups so i'm going to

00:07

be pretty pedantic in this lecture and

00:09

the reason why is

00:11

doing this correctly is a necessary

00:12

condition for

00:14

starting as successful or building a

00:16

successful startup

00:18

so the acronym kpi stands for key

00:20

performance indicator

00:22

if you google around for it there are

00:24

actually many definitions of what this

00:26

actually means but for the purpose of

00:28

today for this context i'm going to

00:29

define it

00:30

as a set of quantitative metrics that

00:33

indicate how healthy your business is

00:35

doing so

00:37

this is important because obviously you

00:39

should know what state your business

00:41

is in at all times so setting the right

00:44

kpis and goals will objectively tell you

00:46

if you're doing well just okay or bad

00:51

so nothing keeps you more grounded

00:52

humbled and realistic about where you

00:54

are than a bunch of numbers because if

00:55

you interpret those numbers correctly

00:57

they don't lie

00:59

it'll also act as a feedback mechanism

01:01

for whether your current strategy like

01:03

user acquisition

01:04

building new features launching new

01:05

features and so on and so forth

01:08

are actually working so if you do

01:10

something and things go up that's

01:11

probably good

01:12

if you do something some things go down

01:13

that's probably bad

01:16

and it will and it will not only help

01:18

you prioritize your time but also

01:20

uh course correct so it follows if you

01:21

do this incorrectly if you set

01:23

your kpis and goals incorrectly you can

01:25

direct your startup

01:26

into a bunch of circles or if you do it

01:29

for too long

01:30

on to the wrong path it will lead to its

01:32

unnecessary demise

01:34

so what are the right kpis to set i'm

01:36

going to break this down into

01:38

two pieces primary metric and secondary

01:41

metrics

01:42

and most of today is going to be focused

01:44

on the primary metric

01:47

so every week in startup school we've

01:49

asked you in the software

01:50

to fill out to define your primary

01:53

metric

01:54

and then update its current value by

01:56

definition you can only pick one

01:58

um one primary metric and as the metric

02:00

if you if you had to you'd

02:01

be willing to bet the whole company on

02:04

so why just

02:05

one metric uh it's a way to focus and

02:07

keep things very simple if there's a way

02:09

to get 90 of the job done

02:11

with just one variable that's better

02:12

than having a bunch of variables that

02:14

gets

02:14

let's say 91 of the job done um in this

02:17

case the job to get done is quickly

02:19

determining

02:20

how well your startup is doing so what

02:23

are the characteristics of a good

02:24

primary metric there are four of them

02:27

uh one so your primary metrics should

02:30

quantify how much value you're

02:31

delivering to your customer

02:33

that is you obviously want to build

02:35

something that people want

02:36

now how much do they actually want it

02:39

and users often indicate the value

02:41

through either training

02:42

uh you through money or time

02:46

so revenue is always the best metric i

02:48

pay you a hundred dollars to use your

02:50

product your software i must at least

02:52

value that

02:53

a hundred dollars um active users uh

02:56

using the product once a week

02:57

or uh once a day we call that weekly

03:00

active user daily active user

03:03

is a weaker but another good decent

03:06

indication of whether you're delivering

03:07

value or not

03:09

the second one here is it your primary

03:11

metric must capture

03:13

whether your product is recurring or

03:15

enduring value to your user

03:17

or it should anyway so for example in

03:19

assassto most sas tools use

03:21

mrr monthly recurring revenue

03:24

as their primary metric i commit to

03:26

forking over 100 bucks a month

03:28

continuously every month because your

03:30

product demonstrates to me every month

03:33

um that has value to me another example

03:35

is if you're building an online digital

03:37

daily newspaper then obviously dau daily

03:40

active user is a good one because i

03:42

expect you

03:43

i expect to be delivering content to you

03:45

that is valuable to you every single day

03:47

so hopefully you'll come back every day

03:49

the third one here is your primary

03:50

metric should be a lagging indicator for

03:52

success

03:53

so a common trap that founders do to

03:55

trick themselves is by picking a primary

03:57

metric

03:58

let's say something like email signups

04:00

because one it's easy to move

04:03

but while it may eventually influence

04:06

revenue

04:07

or actual usage it actually doesn't

04:09

represent

04:10

real value the best so the best

04:13

indication is when

04:14

the value has already been delivered

04:15

it's already occurred

04:17

so when someone has already forked over

04:18

their time or money then to use it then

04:21

that is what a ladder

04:22

that's a definition of what a lagging

04:23

indicator is

04:26

so if revenue increases it's because

04:28

more customers

04:29

have already paid for the product's

04:30

value versus a potential customer who

04:32

came to your site

04:33

gave you an email and maybe they'll sign

04:36

up one day or maybe they'll use your

04:38

product one day

04:39

to buy something and lastly your primary

04:43

metric should be usable as a feedback

04:44

mechanism

04:46

that is it helps you prioritize

04:47

strategies and make decisions quickly

04:50

in a startup one of the key things to be

04:53

to being successful in getting past

04:55

product market fit stages to iterate

04:57

very fast right

04:58

so while you want it to be a lagging

04:59

indicator you also don't want it to lag

05:01

too much so for example a lot of people

05:03

pick mau monthly active user

05:07

but this is but but uh this is usually

05:10

not a great metric because it takes time

05:11

to understand the impact of movement

05:14

um especially in a startup this early as

05:16

in your startup

05:17

and so many things can happen within a

05:19

month

05:21

and also another reason why i don't like

05:22

mau generally is because

05:24

if your user only comes back once a

05:26

month they only value something that

05:28

you're building once a month i really

05:29

question actually if you're solving a

05:31

real problem

05:33

all right so you may have guessed from

05:35

me talking about these four characters

05:37

characteristics of a primary metric that

05:39

there are really

05:40

two primary metrics to pick from so one

05:42

is either revenue

05:44

or active users ideally you're picking

05:47

revenue because nothing tells you more

05:49

about delivering real value than people

05:51

forking over handing over

05:52

real hard earned dollars to you um

05:55

and even better is picking revenue that

05:57

people keep giving you over and over and

05:59

over again like monthly recurring review

06:01

mrr

06:02

it's the best test for whether people

06:03

really want what you're making

06:06

so that being said some people do pick

06:09

revenue but a common trap they fall into

06:11

is that they don't actually get paid and

06:14

usually i hear something to the variant

06:16

of

06:16

oh i'm going to get these i have these

06:18

1000 users not paying me anything

06:20

i just want to get their feedback and

06:22

see how they're using the product and

06:23

make a little bit better and then

06:25

eventually

06:26

i'll get them to pay or the next 1000

06:28

users i'll get them to pay

06:30

that's a trap because free users will

06:33

give you different types of feedback

06:34

than users who are actually paying you

06:37

um paid users are just more serious

06:39

about the product and hopefully will be

06:41

more serious about giving you feedback

06:45

so i urge you to just get paid

06:48

[Music]

06:49

all right so what are reasons why

06:52

so kevin in an early lecture said 99 of

06:55

you

06:55

should actually use revenue as your

06:57

primary metric so what are reasons why

06:59

you should

07:00

um consider active users so one main one

07:03

is because

07:04

you because building a large audience is

07:07

actually a prerequisites to

07:08

modernization

07:10

so an example of this is if your

07:11

business model is advertising based like

07:13

a facebook or google

07:14

then yeah you need millions and millions

07:16

of users coming back to your site every

07:17

day

07:18

before you can actually get brands and

07:20

people to buy ads

07:23

and so in this case active users is

07:25

actually a reasonable proxy for revenue

07:27

because eventually when your startup

07:28

starts making money it's usually

07:30

just revenue is just a multiple of your

07:32

active users

07:34

another reason is also but much much

07:38

more

07:38

much more rare is if you have very

07:41

strong network effects that is if you're

07:43

like a marketplace that requires

07:45

tons of users to just get the flywheel

07:47

going and grow

07:48

then maybe that's a reason for you to

07:50

focus on active users today

07:52

versus uh revenue and then just do

07:54

revenue later down the road

07:57

now that being said if you're using

07:58

active users as a metric

08:00

it's important that you define user

08:02

appropriately

08:04

i hear often i ask okay how many users

08:07

do or

08:08

what's your primary metric uh active

08:09

users how many users have i have 100

08:11

users

08:12

what does users mean in that situation

08:14

sometimes to people it means

08:16

100 users that just signed up and gave

08:17

you an email sometimes it means

08:19

100 users that signed up and start using

08:22

a product and come back every day for

08:23

about 10 minutes a day

08:24

which is by far much better than just

08:26

people just

08:27

like little dabbling on your site right

08:29

so you really need to get that

08:30

definition correctly and don't trick

08:32

yourself by just saying users and get

08:33

having a really

08:34

easy definition of users

08:39

another another example of users where

08:41

it's not exactly users is if you're in a

08:43

marketplace

08:44

and there are two types of customers or

08:46

two types of users so a good example is

08:48

airbnb

08:49

who are your two users you have not just

08:51

the guests

08:52

but you also have the hosts so what are

08:55

you to do how do you

08:56

how do you pick just one well you pick a

08:58

value that actually represents

09:00

um them both getting value so in thermis

09:03

case it would be knights booked

09:05

right another example is uber so who are

09:08

your two users there you have

09:10

riders and you have drivers and so an

09:13

example of a primary metric you could

09:14

pick there is

09:16

weekly trips okay all right

09:19

so uh there are always exceptions to the

09:22

rules

09:23

and there are there is one exception in

09:25

which

09:26

your primary metric is neither revenue

09:28

or active users

09:30

and that is if you run a biotech a hard

09:33

tech business

09:34

and you're still trying to figure out

09:36

whether the science or tech

09:38

is actually going to work can you

09:40

actually build a product

09:41

um and another definition of this is uh

09:44

for our biotech card documents

09:46

businesses it it often takes a lot of

09:48

time and money to get

09:50

your first product to market so what's a

09:53

founder to do especially you have little

09:54

little funding

09:56

so there's two answers to this one is if

09:59

there are no regulatory issues to doing

10:01

sales pre-product

10:02

you should actually do the same as

10:03

everyone else it should be most likely

10:06

revenue your primary metrics should be

10:08

revenue in the form of paid contracts

10:10

lois

10:11

poc's proof of contracts proof that if

10:14

you build it they will actually come

10:16

now if you are in a space with

10:18

regulatory issues meaning you can't sell

10:19

it at all

10:20

um without having to go through like fda

10:22

or some kind of body like that

10:25

then your product primary metric is

10:26

actually less quantitative per se

10:29

and more of a binary thing so it's about

10:32

figuring out the technical milestones

10:34

that you that you need to demonstrate to

10:36

mitigate the risk

10:38

of whether the drug or tech is working

10:40

so if you have to think about experience

10:42

to prove this out you can ask a question

10:45

like what are what minimal things i need

10:46

to do to truly answer

10:48

the question of whether this works or

10:49

not so if

10:51

you fall into this category i urge you

10:53

to actually just go watch

10:55

these two lectures uh they're actually

10:56

firesat chat chats i did

10:58

last startup school with elizabeth and

11:00

eric

11:01

elizabeth is an expert in biotech and

11:04

eric is an expert in hard tech and they

11:06

actually go through

11:07

deep deep dive into how do you think

11:09

about your goals and how do you think

11:11

about your milestones

11:12

and what metrics to actually track all

11:15

right

11:16

so people have referred

11:19

to the primary metric as a north star

11:22

metric and i actually don't like the

11:24

term north star because it kind of

11:26

people have interpreted as something you

11:28

just focus on this one metric and then

11:30

ignore

11:31

everything else um but like i said

11:34

earlier there's

11:34

no metric that actually tells the story

11:37

that tells 100 of the story maybe 90

11:40

but not 100 and so sometimes founders

11:42

fool themselves

11:43

by literally only tracking their primary

11:45

metric and nothing else

11:47

um so a common example is just looking

11:49

at user growth and just ignoring

11:51

retention completely

11:52

but retention is obviously just as

11:54

important to user growth as it is

11:56

as is a new user acquisition

11:59

so one suggestion i have is to

12:02

select a set of three to five other

12:04

metrics secondary metrics to pair with

12:06

your primary metric

12:08

this gives you a good 360 degree

12:10

overview

12:11

of the health of your company so there

12:14

are a ton to choose from so many choose

12:16

from

12:17

what you choose is actually very

12:18

dependent on your business

12:20

next week we're going to have two

12:22

lectures on these sorts of metrics

12:24

for consumer i'll be giving one on

12:26

consumer startups and

12:28

another yc partner i knew will be giving

12:30

one on b2b companies

12:32

and so we'll deep dive into metrics

12:34

these metrics

12:36

next week uh the key here though

12:39

is just picking a few right uh at most

12:42

fives

12:42

three to five close to probably the

12:44

three um you don't want to boil the

12:46

ocean and pick everything

12:48

it's totally fine to track all this kind

12:49

of stuff but it's really not a good idea

12:52

to optimize too many at once to really

12:54

just suffer from analysis paralysis

12:56

all right so a common question i have

12:58

when i say you should

13:00

what is your primary metric you should

13:01

set one is well what if i haven't

13:03

launched yet

13:04

well obviously metrics don't matter if

13:06

you don't know what the problem you're

13:08

solving is you don't even know who your

13:09

customer is yet you should really just

13:10

focus on that first

13:12

um you'd be really putting the cart

13:14

before the horse by

13:15

worrying too much about this kind of

13:17

stuff that said

13:18

once you get to the point where you're

13:20

building the product it's really a good

13:22

idea to get this down even if you

13:24

haven't launched yet

13:25

by at least defining your primary metric

13:27

you'll be able to think about who your

13:28

user really is

13:30

uh you get everyone on the same page on

13:31

who you're targeting and even you can

13:33

hypoth

13:33

use the metrics and goals to hypothesize

13:36

on how you might get your first few

13:37

users

13:39

and trust me nothing is more motivating

13:41

than staring down the barrel

13:43

of zero users and zero dollars of

13:45

revenue

13:46

for weeks on end you're going to get

13:48

very antsy about launching

13:49

um very quickly and that's that's

13:52

actually the effect you want

13:54

all right so i'm going to go into how do

13:56

you set goals

13:58

for your primary metric for for your

14:00

kpis

14:01

so paul graham actually wrote a great

14:03

essay a few years ago

14:05

called startup equals growth and

14:06

explains why startups should focus on

14:08

growth and i really urge you to go read

14:09

it

14:10

and this section of this lecture draws a

14:12

lot of insights from it

14:14

the goal of your startup is to grow your

14:16

primary metric by doing this it does two

14:18

things

14:19

it proves that you're making something

14:21

lots of people want

14:23

and second it proves you're making

14:24

something that has a possibility of

14:26

reaching and serving

14:27

all those people each week your goal

14:30

should actually be to set a weekly uh

14:32

growth rate

14:33

now we use weekly increments because

14:35

startups early on

14:36

need frequent feedback from their users

14:39

to tweak what they're doing

14:41

but also we use weekly growth right

14:42

because it helps to divide

14:44

up the progress you need into doable

14:46

chunks so

14:47

say your goal in a couple months is to

14:48

get 10 000 daily active users

14:51

which requires growing new users let's

14:53

say 10 week over week

14:55

to grow 10 this week may amount to

14:57

actually just getting a hundred new

14:59

users

15:00

which is a different problem to solve

15:02

than trying to get

15:03

10 000 new users right

15:07

you should be focusing on what's

15:08

directly ahead of you in that week

15:10

do things that don't scale today if

15:12

that's actually the best way to get

15:13

those hundred users

15:15

and don't worry about the eventual goal

15:17

of ten thousand

15:19

too soon so naturally the next question

15:22

is how fast should i grow

15:23

what should this rate actually be

15:27

well there's no good formula there's no

15:29

right formula for this

15:30

but one angle that we could tackle it

15:32

from is

15:34

looking at good startups and seeing how

15:35

fast they were growing in the beginning

15:37

stages

15:38

of their life so i actually went back

15:41

and i looked at

15:41

the good startups who pitched in recent

15:43

yc demo days uh so these if you think

15:45

about these startups they were

15:47

three months prior they were all in the

15:49

phase that you are probably in today

15:51

and it turns out the growth rates range

15:53

anywhere from

15:55

twenty to two hundred percent month over

15:57

month but clustered more closely to

15:59

twenty to fifty percent month over month

16:00

which you could back up

16:02

back it out it amounts to about five to

16:04

ten percent week over week

16:05

um and so this chart uh just to explain

16:07

it real quickly

16:08

uh the left-hand column is the weekly

16:11

growth rate and then

16:13

these are the equivalents that you need

16:14

to grow by month and then what the

16:16

multiple is by year

16:18

so this is actually in line if you read

16:20

that essay pg wrote a few years ago

16:22

which he said a good growth rate during

16:24

yc is five to ten

16:25

five to seven percent a week if you can

16:27

hit 10 a week you're doing exceptionally

16:29

well

16:29

and so this is the green area which

16:31

we've seen consistently actually

16:33

um in the recent batches of yc

16:36

so growth is a little hard to grok but

16:38

if you look at this chart you'll see

16:39

that how small variations

16:41

in weekly growth rates can make a huge

16:43

difference on the monthly and yearly

16:45

time horizon you also get the sense that

16:49

to get big

16:50

fast it actually seems doable if you

16:52

have something people want

16:54

on the flip side if you only manage one

16:56

percent uh weekly growth

16:58

it's a sign you haven't figured out

16:59

things yet it doesn't mean that you have

17:01

a horrible business you can run a great

17:03

small profitable business growing one

17:04

percent

17:05

week over week but it's not a good sign

17:07

that you have a startup with a billion

17:09

dollar potential

17:10

so you should think about that trade off

17:11

there and what you really want out of

17:12

your business

17:13

uh if you're growing uh at that rate

17:17

um that said the main thing uh in terms

17:19

of setting

17:21

your goals is is to think for yourself

17:23

uh is to define your own goal based on

17:25

not what others are doing but what you

17:26

think is ambitious

17:28

and achievable based on the product

17:30

you're building

17:31

so you knew your users and business

17:33

better than everyone else

17:34

what does success look like like for you

17:36

and what does being on track look

17:38

like to you so here are some general

17:41

guidelines

17:42

uh when defining a goal all right

17:45

first um if you're solving a real

17:47

problem in a large market

17:48

then that means there's a ton of latent

17:50

demand out there uh

17:52

people will use of just about anything

17:53

to use your product even if it's half

17:55

broken half baked or just solves a bit

17:57

of their problem

17:58

which means that startups usually have

18:01

fast initial growth

18:03

that said where you are today matters so

18:06

if you have a ton of users and a ton of

18:07

revenue

18:08

you will probably know that at that

18:09

volume as a volume increases

18:11

what you need every week to grow gets

18:14

harder over time

18:15

so again most startups they grow very

18:17

quickly and then over some time they

18:18

kind of

18:19

the growth rate kind of slows down a

18:21

little bit the second one is time to

18:24

sale so

18:24

when you try to set your goal you need

18:26

to consider how long it takes to acquire

18:28

user and make a sale

18:29

so for a consumer startup generally you

18:31

have an app or a website

18:33

i show up to it i look at it i see if i

18:36

want it

18:36

and then if i do bam i buy it or i sign

18:39

up for it

18:40

and so it's instantaneous for an

18:42

enterprise startup where

18:43

you're actually probably going through

18:44

some red tape um you have a bunch of

18:46

stakeholders you have to deal with

18:48

um and it just like you you can show up

18:50

to the company and they're not going to

18:51

even buy it right away because you're

18:53

maybe not even talking to the right

18:54

person

18:55

so it might take some months to actually

18:58

get your first sale so you'll have to

19:00

take that in account

19:02

over time this time to sale

19:05

should actually decrease over time like

19:06

good enterprise startups

19:08

that time sale goes from months to

19:09

hopefully days um

19:12

uh if not hours and so it shouldn't

19:14

impact your growth rate

19:15

in the future but in the near term it

19:17

actually might

19:18

uh third is you really want to focus on

19:21

organic

19:21

versus paid users or paid growth in the

19:23

beginning organic means they discover it

19:25

through word of mouth

19:27

basically you're not paying for the user

19:29

they kind of just

19:30

may be searching for it and using it

19:31

themselves i think in the early days

19:33

using paid users is actually cheating

19:35

growth

19:36

and you should avoid it as much as

19:37

possible and finally

19:40

because your startup startups equals

19:41

growth you should focus on exponential

19:43

goals and not

19:44

linear goals all right

19:47

so in terms of picking the goals i think

19:49

there's two ways to do it

19:51

one you can just pick a growth rate

19:55

and then pick up growth rate that you

19:57

can you think you can hit and if you hit

19:58

it great you probably should change it

20:00

if you're hitting it consistently

20:01

to something higher if you don't if

20:04

you're not hitting it then you should be

20:05

a little bit alarmed and you should

20:06

figure out

20:07

why another way to do it is time box an

20:10

absolute goal

20:11

so what i mean by for that is say for

20:13

for the purpose of startup school

20:15

at the end of startup school how many

20:17

active users or how much revenue do you

20:19

want to have

20:20

how what what would it look like what

20:21

would something meaningful look like at

20:23

the end of

20:24

10 weeks then go back out your weekly

20:27

growth rate

20:28

and then go week to week figure out the

20:30

obstacles and how you should

20:31

hit that hit that weekly goal in the

20:34

beginning

20:35

if you're somewhere close to zero users

20:37

today uh often you'll get something

20:39

higher if you

20:39

higher if you do this method than five

20:42

to seven percent week over week

20:45

tracking progress so metrics and goals

20:47

obviously don't mean anything if you

20:49

don't leverage them use these as a

20:52

motivational tool so one way to do this

20:54

is get a piece of paper draw a

20:56

forward-looking graph

20:58

of what the growth you want to hit in

20:59

the next 10 weeks print it out and put

21:01

it everywhere put on top of your desk

21:03

put on the bathroom mirror put on the

21:04

fridge

21:06

and update once and once once a week

21:07

this is in fact what airbnb founders did

21:09

in the beginning

21:11

and if they hit the numbers great if

21:13

they did not and that's all they would

21:14

talk about and so i would follow

21:16

something to

21:17

like this now you want to leverage your

21:20

parametric and goal

21:21

to help you prioritize your time week

21:24

over week so week to week you should be

21:25

stack ranking all the ideas you have of

21:27

how to grow it

21:29

and make a good guess on what's going to

21:31

have the biggest impact for the next

21:32

week to meeting your goal

21:33

and then choose accordingly occasionally

21:36

you won't hit your goal for the week

21:39

we can dream that our growth will be

21:41

flawless and look like this

21:43

but in reality in the beginning it

21:45

always looks something like this

21:47

it's okay if you don't hit your goal one

21:49

or even two weeks in a row

21:51

as long as you understand why you should

21:53

be always asking yourself what is the

21:55

biggest obstacle in my way

21:57

of hitting my weekly target how do i

21:59

overcome this and be obsessive of this

22:02

uh if you don't know the answer then the

22:04

answer is go

22:05

talk to more users and

22:09

don't spin in circles i'm trying to

22:10

figure it out yourself

22:12

a good startup idea will keep growing at

22:15

some point

22:16

so not hitting uh your weekly targets

22:19

week on end

22:20

uh will maybe just help inform you

22:22

you're not working on the right thing or

22:24

even the right idea finally to end as

22:27

you already know our startup school

22:28

software asks you to set your primary

22:30

metric and goals

22:32

uh it is important to be honest about

22:34

where you are

22:36

and one of the best ways to do that is

22:37

to fill this out every week

22:39

we've given you the software to do this

22:41

very easily it is

22:43

not for us i promise you it is for you

22:46

to use and get in the habit of doing it

22:49

we hope you fill us out throughout the

22:51

course and moving forward even after the

22:53

course you keep doing it it's a good

22:55

habit to have

22:57

i guarantee you if you're not already

22:58

doing this just adding this one simple

23:01

thing to your workflow

23:02

is going to help you and change things

23:04

dramatically

23:06

all right that's it we'll next have ilya

23:08

from segment thanks

23:30

you