00:09welcome to metrics for Consumer startups
00:12in our video on metrics for B2B startups
00:15we talked about net dollar retention and
00:17gross margin metrics that are most
00:19important for B2B companies now we're
00:21going to dig into metrics that are
00:22particularly important for Consumer
00:24companies I was previously the founder
00:26of a consumer online bank called monzo
00:29it's now up to about 8 million customers
00:31in the UK and I also worked at grouper a
00:34YC company that built a a group social
00:36club and dating app so two consumer
00:39companies I've also worked with dozens
00:41of consumer companies here at YC for
00:43Consumer companies growth headline user
00:46growth is the most obvious metric that a
00:49lot of people track the reason is often
00:52monetization comes later um you might
00:55have to build up some kind of network
00:57effect or viral coefficient things will
00:59go into later a good growth rate is 15%
01:02month over month at that rate you'll 5x
01:06your user base every year 10% monthly
01:08growth for Consumer company is okay it
01:10means you'll approximately triple your
01:12user base every year 5% a month or lower
01:15is unlikely to reach breakout success
01:18I'm afraid but growth is much more
01:20complex than simply headline user
01:23numbers we'll split first of all into
01:25organic versus paid growth so organic
01:27growth is really anything you don't pay
01:30and Founders running consumer companies
01:32often neglect this we got to a million
01:35customers at monzo before we spent any
01:38money on Direct marketing or advertising
01:40and really we did this in two ways
01:44virality and network effect it's worth
01:47pausing on these because they're so
01:48important virality is the idea that one
01:52user using your product introduces it to
01:55other users somehow in the use of your
01:57product whereas Network effect is the
02:00idea that the product gets better uh
02:03that more nodes in the network exist
02:04I'll give you an example from Facebook's
02:06early days so when you took a a picture
02:09on Facebook and uploaded it you prompted
02:12to tag your friends even if they weren't
02:14already on Facebook and those friends
02:17would get an email saying hey someone's
02:18tagged a photo of you sign up to view
02:20the photo that's a viral mechanic by
02:23uploading the photo and tagging your
02:25friends you spread it to New users
02:27another example might be Wordle so when
02:29people play Wordle they post their score
02:32on social networks you see those little
02:34green and and gray dots that's a signal
02:36for everyone else there they're playing
02:38uh this game and it attracts new people
02:40into the game so that's virality a
02:44network effect comes from meta's law the
02:46value of the network is the square of
02:48the nodes in a network and what that
02:49basically means is the more people who
02:52use this thing the more valuable it gets
02:54and so WhatsApp is a great example of
02:56this if you're the only one on WhatsApp
02:58you know you can't message with anyone
03:00it's pretty useless but the more people
03:03you add to the product the more people
03:05you invite the more useful it gets for
03:07you you can message all these people
03:09around the world for free and so the
03:11best consumer companies incorporate both
03:14virality and network effect which are
03:17different concepts but very closely
03:18related in order to grow their user base
03:21organically so let's think about your
03:24product How Could You incorporate both
03:27of these things virality and netw effect
03:30first of all for virality what are the
03:32sharable moments what are the points at
03:34which you you know you accomplished a
03:37something new in the product you've you
03:39know dualingo you've reached a new level
03:41something you want to brag about or
03:42Wordle you've completed it in two tries
03:44or something uh what are those points in
03:47the product uh where people naturally
03:50are inclined to share it and then how
03:51can you make it really easy with all of
03:53those sharing prompts that um both IOS
03:55and Android offer the second is Network
03:58effect how could my product get better
04:02than more people who join you have to
04:04shift from thinking about your product
04:05as a single player journey to a
04:07multiplayer Journey so for something
04:08like monzo which was a Neo Bank the
04:10network effect with things like being
04:12able to send and receive money really
04:14quickly within the bank so we've built a
04:16venmo style um money transmission within
04:19the bank you could also open joint
04:21accounts or have big pots for groups of
04:24people who are going on a holiday and so
04:25what we'd often see is a group of six or
04:28seven people go on a holiday
04:30the St only three people have monzo and
04:32by the end of the holiday uh all of the
04:34group have been bullied to to sign up to
04:36monzo so they can jointly manage their
04:37expenses that's a network effect working
04:40on these Network effect and viral Loops
04:43will pay back every day for the rest of
04:45the life of the company where with ad
04:47spend you spend the money one day
04:49tomorrow it's gone you've got to keep
04:51spending that money to keep acquiring
04:52users viral loops and network effects
04:55pay back forever so every one or 2%
04:58extra you can op optimize in your viral
05:00loops and your network effect will pay
05:02back for the rest of the life of the
05:03company paid referral schemes are a sort
05:06of interesting blend it seems like
05:08member get member where you know if you
05:10refer a friend you get $5 and they get
05:12$5 or maybe with Uber you get a free
05:13ride they get a free ride I would treat
05:15this as paid acquisition actually you're
05:17spending money to acquire customers and
05:19if you stop spending those customers
05:21won't won't appear there are two things
05:23to watch out for if you're doing this uh
05:25these paid referral schemes first of all
05:27is cannibalization this is the idea that
05:30people would have referred their friends
05:32anyway and now by paying them you're
05:34simply paying for users that would have
05:37signed up for free so that's
05:39cannibalization and you can test that by
05:41only running your paid referral schemes
05:42in certain parts of the country or
05:44certain cities or turning it off for a
05:46period of time and seeing what natural
05:48organic rate of referral you get the
05:50second thing you need to watch out for
05:52is fraud honestly people somehow milking
05:55your referral scheme so I had a friend
05:57who in the early days of um of zip car
06:00just set up a bunch of cheap Google ads
06:02that bid on zip car and then would
06:04redirect them straight to Zipcar and
06:05milk the referral scheme and so he drove
06:07free Zip Cars for a year and then got
06:09banned for life from zip car so there
06:11are always people doing this you've just
06:12got to watch out for it it's just an
06:14annoying cost okay so we talked about
06:17organic growth and particularly Network
06:19effect and and viral Loops now we're
06:21going to talk about paid growth so this
06:23is the idea that you you know you you do
06:26a pay-per-click campaign maybe you do
06:28some TV advertising ing or advertising a
06:30newspaper or whatever the first thing is
06:33you have to have good tracking set up
06:35you need to know where every user is
06:37coming from did they come from your
06:39Facebook ads or your Instagram ads or
06:41that TV ad the simplest way for
06:42pay-per-click ads is that UTM referr in
06:45the URL if you can't get that or I know
06:47recent changes to iOS has made it much
06:49much harder to track you should just ask
06:51your users when they signed up where
06:53they heard about you you so you have to
06:55measure this paid versus free and you've
06:56got to know where they came from you can
06:59get really convoluted with you know
07:01first touch attribution last touch
07:03multi-touch attribution
07:05it's it's often way more hassle than
07:07it's worth simply using those UTM
07:10referrals or asking your customers where
07:11they came where they heard about your
07:13your business is probably the best way
07:15for most of you so understanding where
07:17each user came from and then how much
07:19you spent on each Channel lets you
07:22understand your customer acquisition
07:23cost how much you spent for each channel
07:26to acquire each User it's important that
07:29you track this per Channel and then
07:32crucially record that in your database
07:35where each user came from record it
07:36forever because you can then monitor the
07:39performance of those customers over time
07:41back at monzo we found a particular um
07:44money-saving blog that was very very
07:46cheap to get users from they sent us
07:48hundreds and hundreds of users uh when
07:50we were eventually paying for
07:51advertising uh it seemed very very cheap
07:54but the customers it turns out were
07:55deeply unprofitable they would load
07:57money onto their monzo card they'd go on
07:59and then they instantly just go and take
08:01,000 out from the ATM and just spend in
08:03cash and we couldn't understand why this
08:05was for a very very long time we
08:07researched it but the the net result of
08:09it was we shut down that advertising
08:11Channel even though the cost of
08:13acquisition looked really really good
08:14for that channel the lifetime value of
08:17those customers was negative the the
08:19revenue and the profit they generated
08:20was negative for the company because
08:22they generated so much cost the second
08:24thing for to watch out for for for CAC
08:26for customer acquisition cost is you
08:29have to measure it to an active
08:31monetized retaining user whether
08:34whatever that's defined as for your
08:35business it might be a subscriber for
08:37monzo it was a weekly active user you
08:39have to measure it to get to a good user
08:41not simply a sign up because you might
08:43have 80 or 90% drop off rate in your
08:46first week for example you've got to
08:48track what does it cost us to generate a
08:51user who sticks around for the longterm
08:53who performs and retains like a like one
08:56of our good users that's your customer
08:58acquisition cost the best consumer
09:00companies have a split of organic versus
09:04paid growth of something like north of
09:0780% organic to 20% paid even 100%
09:11organic to 0% paid for some of the
09:13absolute best consumer companies like
09:15Facebook and and WhatsApp in the early
09:17days a 50/50 split between paid and
09:19organic is okay anything below 50/50 for
09:23any period of time honestly is pretty
09:24worrisome that's because if you're
09:26relying too much on the big ad platforms
09:28like Google and meta to increase your
09:32growth you're basically just going to
09:33have to pay them more money and as you
09:35pay them more money to acquire more
09:36users the cost to acquire each user goes
09:38up and these ad platforms are really
09:40well tuned to extracting the maximum
09:43value from you and your competitors as
09:45possible and so what happens is all
09:46these competing companies bid and bid
09:48and bid to get more and more and more
09:50customers and see they're making less
09:52and less profit on each customer because
09:54they're requiring more and more to to
09:55acquire them and at the end of the day
09:57only Google and meta really win in that
10:00battle You all take your margins to zero
10:04and Google and meta end up making all
10:05the money that's why it's so dangerous
10:08to be so reliant on paid growth
10:10similarly a big platform shift like the
10:12changes to uh to the iOS platform to the
10:15broke advertising tracking just wipes
10:17out half of your business I've seen so
10:19many companies overnight just die from
10:22changes that made were made to the big
10:24advertising platforms so please please
10:26please focus on organic growth and
10:29that's viral loops and network effect so
10:31to finish up on paid growth I've not
10:34seen any great consumer company get to
10:37True scale where more than 50% of their
10:40signups are coming from organic JS and
10:42I'm sure someone out there will watch
10:44this and find an example but
10:45overwhelmingly the great consumer
10:47companies that get to IPO scale to get
10:51to really big scale um have optimized
10:54their viral loops and their Network
10:55effects to get the majority of their
10:57growth from organic channs so now we're
10:59going to dive into unit economics this
11:01is the idea that you can measure on a
11:03per customer level how much revenue they
11:05generate minus the variable or
11:08incremental cost associated with serving
11:10that warn customer so let's take monzo
11:12an online bank when I was there just a
11:14few years ago these numbers aren't
11:16current each customer generated about
11:18$50 in Revenue but we had a lot of costs
11:21to serve each customer so for example
11:24sending out a new plastic card and a
11:26replacement plastic card whenever they
11:28were lost was a variable cost the cost
11:30of customer service so the more each
11:33customer contacted customer service the
11:35more cost that would generate any fraud
11:37on the account or any transaction fees
11:40to to make a bank payment or wire fees
11:42all of those were cost that monzo would
11:45bear and then we track those on a per
11:47customer basis this is very very
11:50important because you can then
11:52understand not just how profitable or
11:55unprofitable the entire customer base is
11:57but for each sing single customer and
11:59each cohort of customer how they
12:02performed so for example we discovered
12:05that customers at monzo who traveled
12:07abroad very often would earn a lot more
12:09revenue or customers who came in Via a
12:12certain advertising Channel I mentioned
12:14earlier might take money out of ATMs and
12:17incur massive costs there or certain
12:18kind of customers might contact customer
12:21service very very regularly and incur a
12:23lot of cost there and so this really
12:26matters in an operationally complex
12:28business where you've got thin margins
12:30it's really important that you
12:31understand on a per customer basis how
12:33much revenue each customer generates and
12:35then how much variable cost each
12:37customer generates and you can optimize
12:40it so you can start reducing the cost or
12:42you can start investing in the
12:44advertising channels that bring the
12:46kinds of customers who tend to generate
12:49higher revenue and lower cost so rather
12:51than tracking these on a a really broad
12:54basis the more granular you can get on
12:56these costs the better so quick note
12:58things that aren't included in variable
13:00costs are those fixed costs the things
13:02that if you doubled or tripled your
13:03customer base really those costs
13:05wouldn't change at all so those might be
13:07engineering salaries or the rent on your
13:09head office those are the things that
13:11you deduct afterwards to get to your
13:14total profit line at the bottom so for
13:16uni economics you really want to take
13:17your Revenue minus your variable costs
13:20those are costs that vary as you
13:23increase the number of customers and as
13:25we talked about in the video on Gross
13:27margins scaling negative unit economics
13:30is very very dangerous in the early days
13:32of monzo we were AT3 or4 per customer
13:37per year and we scale to half a million
13:39customers before we fix them earn a lot
13:41of capital so beware pay attention to un
13:44economics and try to get them positive
13:46before you scale up your user base okay
13:48next retention we talked about retention
13:51in our B2B video and particularly net
13:53dollar retention it's pretty easy in uh
13:56B2B SAS companies when you have a
13:58recurring contract with a paying
14:00customer to know whether they've churned
14:01or not with consumer it's often a little
14:04harder especially when they're not a
14:05subscriber but they just periodically
14:07use your product like Airbnb and the
14:09complicated question is what is the
14:12right period to measure to tell if
14:14there's still an active customer or not
14:16should they be using a product every day
14:18or every week or month or even year so
14:20for Facebook an active user might be
14:22checking Facebook once or multiple times
14:24a day but with Airbnb you might only
14:27book a vacation home every six months
14:30and so for your business it's really
14:33important to think for a successful
14:35customer so want someone who really
14:36likes the product how often would they
14:39typically be using this service so for
14:42monzo we chose one transaction at least
14:44one financial transaction every week it
14:46can be hard to measure retention in
14:48consumer startups especially when you've
14:50decided that the relevant usage period
14:53might be months or even years bit like
14:55Airbnb or an airline booking so what
14:58some companies figured out is you can
15:00look for what's called a magic moment
15:04this user Behavior or activity that is
15:07correlated with or predicts long-term
15:09retention so the way you go about that
15:11is you analyze a group of your best
15:13users and compare their behavior how
15:16they use the product to the rest of your
15:18customer base so a famous example was
15:20Facebook they figured out that if you
15:23added seven friends in your first 10
15:26days of usage you'd overwhelmingly go on
15:28to become long-term happy users of
15:31Facebook whereas if you didn't add
15:33friends in your first 10 days it's much
15:35more likely you churn off for monzo we
15:37had something similar because we had a
15:39bank account but also this venmo like
15:41functionality built into the product
15:42where you could send and receive money
15:44customers who added three friends from
15:47their phone book to their monzo account
15:48so they could send and receive money as
15:50they signed up retain something like 20
15:52percentage points better than users who
15:55didn't have any friends in their first
15:57few days so once you figured out what
15:59your magic moment is might be adding
16:01friends like Facebook or monzo or it
16:04might be booking a stay and having a
16:05fstar experience as with Airbnb you can
16:08try to re-engineer your product
16:10unboarding to make sure that as many
16:13users as possible Hit This Magic Moment
16:16as soon as possible so for Facebook it
16:18was getting that ad friend uh popup
16:21inside the signup flow making sure that
16:23everyone who joined Facebook was
16:26prompted to add friends and so for your
16:28product think about what that magic
16:29moment could be and how you can
16:31re-engineer the signup flow to get as
16:34many users as possible hitting that
16:36point there is a mistake here and then I
16:38definitely uh fell victim to this it's
16:41getting too fixated on that precise
16:44definition of that magic moment for
16:45Facebook it probably didn't matter that
16:47much whether it was seven friends or
16:49eight friends or six friends simply
16:51finding that kind of Tipping Point that
16:53looks about right from your metrics and
16:56then agreeing with everyone that that's
16:57what you're going to optimize for rather
16:59than being really pedantic about the
17:01exact definition it's probably the best
17:03way to go the next thing we're going to
17:04talk about for Consumer startup metrics
17:06is the net promoter score again this is
17:09a three-letter acronym NPS that sounds
17:11pretty complicated but it's pretty much
17:13a measure of How likely you are to
17:16recommend this product to a friend you
17:18ask it to your customers and it's a
17:19scale of 0 to 10 and you basically take
17:24all the people who are promoters and
17:26that's a n or 10 and then you net off
17:29hence the net the detractors that's
17:32people who rate you 0 to six and that'll
17:35give you a score so if everyone is a
17:38detractor all people who vote voted at
17:40zero you'd have a negative 100 100% of
17:43people were detractors if 100% of people
17:46were promoters you had zero detractors
17:48100 minus zero you have a plus 100 net
17:51promoter score another example say you
17:53have 50 promoters 50 scoring 9 or 10 25
17:57in the middle 25 neutral you don't do
17:59anything with those you just ignore them
18:00and then minus 25 detractors those who
18:03people who are scoring at 0 to six 50us
18:0625 your MPS is positive 25 if your net
18:11promoter score your MPS is not extremely
18:13high as a new consumer company your
18:15toast you pretty much have to
18:17re-engineer the product to make it
18:20something that people love the reason is
18:23it correlates extremely well with Word
18:25of Mouth referral so I would argue
18:28positive 50 is pretty much a minimum
18:30Baseline for any new consumer company
18:33you have to be in order of magnitude
18:34better than your competitors mono for
18:36example hovered somewhere between
18:38positive 75 and positive 80 while I was
18:41there I looked up Tesla it's currently
18:43at positive 96 so the best consumer
18:47companies always always have a Sky High
18:50net promoter score and you if you look
18:52at old incumbents people like cell phone
18:54companies or really old Banks their net
18:57promoter scores are often
18:58zero or even negative that's a really
19:01good sign that you could disrupt those
19:03Industries with great customer service
19:05and a great mobile proposition how you
19:08gather this score is extremely important
19:11so you ask this question to your
19:13customers it might be every month or so
19:15with a random sample or at a certain
19:16point in the product but the crucial
19:18thing is that you keep it consistent if
19:21you change the point in the app or the
19:24the point in the customer's lifetime
19:26when you ask this question you're going
19:28to change the values you get you're
19:30going to change the responses and your
19:32net promoter score will jump around
19:34wildly and you won't know whether it's
19:37something you've changed or simply an
19:39artifact of you having changed the
19:40collection method we fell into this trap
19:42at grouper dating website I worked at we
19:45changed the collection method and the
19:47net promoter score decreased by 20
19:49percentage points overnight we couldn't
19:50figure it out so it's crucial you ask
19:53this in the same way of all of customers
19:57at various points in time that you don't
19:59change this and finally on NPS you can
20:01influence this this figure in particular
20:03as a next question after people have
20:05answered whether they recommend this to
20:07a friend you should ask a qualitative
20:10question why why do you like the product
20:13or what's wrong with it and then in
20:15particular you can go and look at all of
20:17the detractors and you can fix all of
20:19the things they don't like and that's a
20:20sure fire way to increase your net
20:22promoter score okay we've covered a lot
20:24of stuff in consumer metrics today let's
20:27quickly recap first of all is your
20:29growth rate at least 15% month-on-month
20:32growth in active users is the target 10%
20:36is okay 5% is probably not going to cut
20:38it for Consumer startup we talked about
20:40organic versus paid growth you really
20:43need at least 50% organic to get to any
20:45kind of scale if you've got more than
20:4750% of your growth coming from Paid
20:49channels I'd suggest in the long term
20:52you have a problem that you need to fix
20:54so focus on virality and network effect
20:57to get the those organic signups humming
21:00if you're not already start tracking
21:02unit economics this is the amount of
21:04Revenue each customer generates minus
21:07the variable cost that each customer
21:09creates if your unit economics are
21:11negative you must fix them before you
21:13scale then we talked about retention
21:16retention needs to flatten off somewhere
21:18if your retention periods are really
21:20long if people only come back every 6
21:21months or year look for a magic moment
21:24in your product that moment where users
21:26finally discover the aha moment and are
21:28more likely to convert to long-term
21:30happy users and the net promoter score
21:33plus 50 as a minimum for great consumer
21:36startups and don't mess around with how
21:38you're measuring net promoter score
21:40you're likely to throw off your metrics
21:41and confuse yourself a final warning
21:44these are all benchmarks your company
21:47and your industry May well be different
21:50you may have off the charts organic
21:52growth and very poor unit economics like
21:55early monzo or you might have really
21:57really good economics and pretty slow
22:00growth that rely on paid ads every
22:03business is different and it's hard to
22:05give cookie cutter benchmarks for
22:06everything so just think about these
22:08metrics and how they might be different
22:10for your business I hope these are
22:11useful in running your company thanks