🔍

The metric that could kill your startup - YouTube

Channel: Slidebean

[0]

churn is cancer for a sas company

[2]

overlooking it may have been the single

[4]

biggest mistake

[5]

our company made and the closest we've

[7]

ever been from going out of business

[8]

churn is a silent killer that slowly but

[10]

surely will eat away your business

[12]

you can cure it if you could catch it

[14]

early on but if you let your churn

[15]

problem go unchecked

[17]

it will kill your starter and we learned

[18]

this the hard way so let me give you

[20]

some context

[21]

into how we found ourselves here between

[23]

2015 and 2016

[25]

our company grew really fast we 5xed our

[27]

subscriptions

[28]

just in 12 months and everybody was

[30]

really excited

[31]

in 2016 we raised the second part of our

[34]

seed round and it was

[36]

the easiest money that i've ever raised

[38]

it was obvious

[39]

that we were on to something with

[40]

slightly and the diligence process for

[43]

what the investors had to do

[44]

was actually pretty smooth there was

[47]

inevitable fomo

[48]

on sniper and i remember one investor

[50]

coming and saying that they were going

[51]

to pass because our churn was too high

[53]

and at that point nobody was really

[54]

paying much attention to churn

[56]

let me show you why you'll see that

[58]

we're adding six to seven thousand

[60]

dollars in new subscriptions every month

[62]

in the second half of 2016 that grew to

[64]

10 to 12 000 subscriptions new mrr

[67]

every month which again got everyone

[69]

really excited at the time

[71]

we were consistently adding 45 000 in

[73]

net mrr and

[74]

moving fast towards the 1 million dollar

[76]

annual run rate

[77]

by the way i'm fast tracking through

[79]

some terms like month return revenue

[81]

ar assuming that you've already seen

[83]

some of our videos

[84]

if you haven't you want to hit pause and

[86]

go watch this specific one sas metrics

[88]

i promise you it'll help you make a

[90]

whole lot more sense of

[91]

everything that's coming in this video

[93]

so the big number of new customers made

[95]

us ignore

[96]

the number of lost customers and we only

[98]

focus on that pretty growing mrr chart

[101]

and it's easy to miss

[102]

what's happening under the hood so let

[104]

me show you what we did

[105]

to fix it the first part was

[107]

understanding

[108]

the why why do people turn

[111]

in general people cancel because they're

[114]

not getting enough

[115]

value out of a subscription not enough

[118]

to justify

[119]

the price period that's really the only

[121]

reason you can talk all you want about

[123]

onboarding aha moments and cancellation

[126]

flows it all comes down

[127]

to the value obtained versus the price

[130]

the more value

[131]

you bring to your customers the more

[132]

money you can get away

[134]

with charging and i get this question a

[135]

lot what should be the pricing of a sas

[138]

product from the retail world we are

[141]

maybe accustomed to thinking about

[142]

margins we mark up the manufacturing

[145]

price

[145]

to generate profits from that in

[148]

software

[149]

it's it's really different because

[150]

software is really cheap to operate so

[152]

we could very much leave money on the

[154]

table if we aren't aggressive

[156]

enough about pricing my take on this

[158]

here is that the pricing should be as

[160]

much as we can get away with charging

[163]

that's simple enough you can find that

[165]

number by testing there's no way to

[167]

guess it and

[168]

user interviews and surveys are not

[170]

going to give you the final answer on

[171]

this they might guide you

[172]

but you need to test and balancing these

[175]

two variables

[176]

the value obtained and the price is

[178]

really some of the most

[180]

complex challenges for sas product

[183]

manager so let's talk about value first

[185]

the value

[186]

that a customer gets from a product can

[188]

be affected

[189]

by number one the onboarding so

[193]

if you are not onboarding your customers

[195]

correctly if you're not explaining

[196]

what the tool does then the customers

[198]

won't extract enough value out of it

[200]

because maybe they just don't know how

[203]

then you have the ux the user experience

[204]

if customers can

[205]

find what they're looking for or if they

[207]

find the experience frustrating

[209]

we are willing to forgive ux hiccups up

[211]

to a certain point

[212]

but after that we'll just cancel number

[214]

three simply how the tool is used in

[216]

their day-to-day are they using it daily

[218]

weekly what's the alternative how much

[220]

of a better experience

[221]

are you giving them versus the

[222]

alternative some examples from our own

[225]

book we

[225]

use intercom for our messaging for our

[228]

onboarding for our emails for our

[229]

support and for our newsletters

[231]

it's over five thousand dollars a month

[233]

for us it's one of our most expensive

[234]

sas platforms but

[236]

we couldn't really live without it the

[238]

alternative doesn't have all the

[239]

features so

[240]

we just have to mix a bunch of different

[241]

tools and it would still cost us

[243]

somewhere in that range

[244]

another service that hasn't logged for

[246]

life is turn module it's the platform

[248]

that you've been seeing in some of our

[249]

youtube videos

[250]

when we refer to sas metrics and the one

[252]

we saw in this video

[253]

we've been tracking our sask apis there

[255]

for years and

[256]

that set of features gives us what we

[258]

need and we don't want to go through the

[259]

process of

[260]

migrating all of this data elsewhere

[262]

it's just a lot of data

[264]

another variable you might have noticed

[265]

here is how hard it is

[267]

to leave so for b2b software you want to

[269]

find something

[270]

that embeds deeply in the work pipeline

[272]

of your customers so that they

[274]

depend on your tool those are the

[276]

companies that thrive look at slack or

[278]

heck look even at microsoft office so

[281]

your focus

[282]

as a founder or as a product manager and

[284]

you're probably going to be both for a

[285]

little bit when you're starting a

[286]

company

[287]

is creating a product that adds a clear

[290]

value to your customers

[291]

lives it can't be just a nice to have

[293]

ideally it should be

[294]

and i can't live without it type of

[297]

relationship

[298]

and there are a lot of ways that you can

[299]

measure this first you want to ask them

[301]

you want to get into conversations with

[302]

your customers

[303]

i do probably 20 or 30 user calls a

[305]

month and that gives me

[307]

this qualitative view as much as those

[310]

calls are focused on helping them

[312]

in whatever they can and whatever i can

[314]

i'm

[315]

taking notes i'm asking which features

[317]

they found or haven't found useful

[319]

and the ones that they've been using

[320]

more frequently for more

[322]

quantitative approach we have a little

[325]

survey that you can use

[326]

you've probably seen it before how

[328]

disappointed would you be

[330]

if you could no longer use slide p

[333]

another

[334]

standardized measure for this is the nps

[336]

service the mps score

[337]

net promoter score the question is how

[340]

likely

[340]

are you to recommend sniping or any

[342]

product to a friend there's a 0 to 10

[345]

meter so any answer below 6 is

[347]

considered bad so that's minus 1 point

[349]

answers between 7 and 8 are considered

[351]

average so that's neutral

[353]

or zero and then the answers that really

[356]

speak passion about the product are

[357]

nines and tenths

[359]

so the sum of your nps score answers

[360]

should be somewhere between 20 and 40.

[362]

40 plus is considered great

[364]

70 plus is considered world class now

[367]

all of these indicators can be

[368]

considered

[369]

predictions for churn what we do is we

[372]

pull

[372]

this data into chart modules so we can

[374]

look at the aggregated data

[376]

based on the nps score and understand

[378]

the likelihood of one of these customers

[380]

churning or even our effectiveness in

[383]

getting people

[384]

from nine to ten to seven and eight

[386]

month per month and how much revenue we

[387]

get from each one of those groups

[389]

last but not least you can also ask them

[392]

why they're leaving you're i mean you've

[393]

seen this

[394]

quite a lot in cancellation flows and

[396]

this is valuable information

[397]

by any means but there's a big catch

[399]

there which is that there's not a lot

[400]

that you can do

[401]

if a user has already cancelled your

[403]

chances of re-engaging them

[405]

at this point are pretty slim so what

[407]

you can use

[408]

is to plug those reasons into your data

[410]

and then study why they're cancelled

[412]

now in our case most of the

[414]

cancellations we were seeing in 2016

[416]

could be summarized as this story i like

[419]

your product

[420]

i just don't need it anymore and that

[422]

was very

[423]

painful to hear the story was consistent

[426]

i can still

[427]

see it in my 2016 nightmares nine or 10

[430]

nps score

[431]

great customer satisfaction i just don't

[433]

need to make presentations anymore

[435]

but that answer will be different of

[437]

course from company to company

[438]

you might be getting complaints i don't

[440]

know on missing features or bugs

[442]

and those i would say maybe they're

[444]

easier or generally easier to tackle

[446]

easier to address because the customers

[448]

are pretty much telling you what

[449]

you want what's driving them away and

[451]

again as long as you're taking the time

[453]

to track it and to make good decisions

[455]

you're good but our cancellation reason

[457]

reflected a deeper a more fundamental

[459]

problem with our product

[460]

this was combined with the fact that our

[462]

second most common reason for canceling

[465]

was pricing people were saying the

[467]

product was too expensive so we figured

[468]

well maybe we're going to play around

[470]

with pricing

[472]

as a solution and boy were we wrong i

[474]

like to call this stage of

[475]

our process half measures because i love

[478]

breaking bad

[479]

and this is a breaking bad reference

[484]

so path measures and pricing now

[487]

remember two variables affect

[489]

churn the value and the pricing we have

[491]

experimented with pricing

[493]

a lot we started with 4.99 then went to

[495]

29 and then to 79

[497]

a month now our biggest lesson from all

[499]

of this is then when converting people

[501]

people are less sensitive to pricing

[503]

that you might think

[503]

one of our biggest breakthroughs as a

[505]

company when we were starting up was

[506]

that we discovered that the amount of

[508]

people that we could convert

[509]

on a 499 a month plan versus a hundred

[512]

and sixty dollars a year

[514]

annual only plan was pretty similar

[516]

conversion rate was only slightly

[517]

affected with the latter pricing and we

[519]

of course

[520]

made a lot of money a lot more money

[522]

with the year subscription now

[524]

our first thesis to solve our turn was

[526]

allowing customers to pay

[528]

to unlock a single presentation rather

[530]

than a subscription so

[531]

we ramp up the price of the subscription

[533]

so that only customers that

[535]

are power users and there are really

[536]

gonna present

[538]

choose the subscription ramp and the

[539]

others can just unlock the presentation

[541]

so we settled for a single presentation

[544]

unlocked at 29

[545]

and then monthly plans starting at 49

[547]

and then 79.

[548]

and i think the thesis is not that bad i

[551]

mean if we look back

[552]

i think the thesis was kind of solid i

[554]

think what would you say

[556]

but boy did it fail at first we saw a

[559]

nice spike of non-recurring revenue

[560]

which

[561]

chet mongol reported separately so we

[562]

could keep track and then while the

[564]

number of subscribers was

[566]

reduced we still saw a decent amount of

[568]

people that were picking

[569]

the subscription option but the real

[572]

problem

[573]

churn would come by this a couple of

[575]

months later

[576]

the new higher pricing spiked churn to

[579]

[580]

people were kind of accepting this new

[582]

pricing to convert

[584]

but they were really really sensitive

[586]

about it and they would cancel

[588]

in a month or two and of course the

[589]

value that this product was bringing was

[591]

not 79

[592]

worth again we thought only power

[593]

presenters would pick the plan but

[595]

that really wasn't the case even worse

[598]

our nps score

[599]

dropped a lot people were suddenly much

[601]

more sensitive to bugs

[602]

and stuff that they couldn't solve in

[604]

the platform as expected

[606]

and they expected it more hands-on

[607]

support than a 79 plan

[609]

now one detail we did not foresee here

[612]

was how much revenue shifted from

[614]

recurring to non-recurring we had to

[616]

change those sales every month and a bad

[618]

month

[618]

would seriously endanger our runway so

[621]

after a failure in that approach we

[623]

considered going with

[624]

a lower pricing we came up with a plan

[626]

called starter

[627]

which was priced at 12 a month and that

[631]

kind of worked also kind of worked the

[634]

churning in the 12 hour month plan was

[635]

better

[635]

under 10 for the first time in a while

[637]

but the next problem

[638]

was lifetime value and cash flow look at

[641]

this math

[642]

on a 79 a month plan with 20 churn your

[644]

average lifetime value per customer is

[646]

395 dollars

[647]

that means the average customer stays on

[649]

the platform for about

[650]

five months not a long time but for cash

[654]

flow you're collecting

[655]

395 dollars in five months now don't get

[658]

me wrong

[658]

that level of churn will kill any

[660]

business but at the time we were

[662]

acquiring customers for about fifty

[663]

dollars

[664]

so we spent fifty dollars and we

[666]

recovered them the first month

[668]

and then we make 345 dollars worth of

[670]

profit within the next

[672]

five months on a 12 a month plan with a

[674]

five percent turn rate your lifetime

[676]

value is 240

[678]

it's less and worse than that it takes

[680]

20 months to collect that money

[682]

it takes you four months to recover the

[683]

fifty dollars you spent bringing the

[685]

customer

[685]

so we had sort of solved the churn

[688]

problem but the pricing created

[689]