Growth vs Retention

Entrepreneurs always ask what the one number they should focus on for raising money. I always say “90 day retention numbers for your acquisition cohorts”. There’s a common view in silicon valley and around the tech sector that growth is the one thing you should focus on. But it’s hard to grow if you are churning your users. And if you are paying for user acquisition, as many startups do in search of growth, then retention/churn becomes even more important.

This issue was highlighted in a Forbes post on Homejoy, which apparently had a retention problem. A former employee said:

Retention was clearly bad, and that’s what killed us

This is a huge conundrum for entrepreneurs who want and need to grow in order to build confidence with investors (both existing and future) and to attract and retain talent.

This all comes back to stepping on the gas before finding product market fit. You might think you have product market fit and so you scale up your hiring, your marketing, your sales, and your capital raising and spending.

But if you can’t retain a healthy percentage of your users past ninety days, you don’t have product market fit yet and all the investment you make in your business is just money down the drain.

So focus first on your 90 day retention numbers and make sure to nail them and prove you have product market fit. Then scale.


Comments (Archived):

  1. David Goldberg

    Fred, any guidance on what you consider a strong 90-day retention for a certain model or vertical?

    1. Manuel Jaeger

      Good question!!

    2. Richard

      For homejoy , the number should be 30 days. If you are not using a cleaning service one a month, you’ve moved on

      1. Jess Bachman

        I’d say 90 days is still better. If the service is only used once a month, you may just be getting the benefit of the doubt for that first month, maybe two.

        1. Richard

          If you calculate hazard rates and survival rates, one can let the data point you to the answer.



  2. LE

    I am wondering how the “90 day” number differs for free vs. paid services.

    1. JaredMermey

      Or service dependent…for something one uses every day 90 days might work, but what about something that is used once a month — should it be longer?

    2. B. Llewellyn Shepard

      Yeah but in either case, you want them to stick around longer.

    3. William Mougayar

      I don’t think it matters, because in both cases, the growth is leaky. it might look slightly better in paid, because users who have paid for something will likely use it a bit longer, but once you discount the paid period cycle and you still see churn, then you’re back to analyzing and solving.

  3. LIAD

    was surprised to see Groupon referenced as a ‘culprit’ in their downfall.over the years we’ve had dozens of high profile companies tell the same story”unsustainable customer acquisition offers -> no repeat usage -> instant churn”annoyed sophisticated investors allowed this co to follow a playbook we all know ends in ruin.

    1. LE

      Groupon at the core could be modified to make sense (as a new business idea). What groupon really does is distribute and highlight businesses that wouldn’t be noticed if not for groupon. At the core. It’s is or should be a lead generating machine.My ex wife ran a coupon advertising book (in the 90’s) that provided value to the advertisers (and she would typically get upward of 70% (iirc) renewals). The coupon book was distributed at college campuses and offered some nominal discount but the important thing was it introduced college students to restaurants, bars, student services and so on. The advertisers paid for the coupon and offered a small discount to the students. The model worked. The hard part was actually the distribution (and collecting the money that was owed). A competitor did the same thing but with coupons to non college students. (They later sold to a large newspaper chain since they covered so many markets). Today there are many people (money mailer for example) that do similar things. The entire idea is to get the business name in front of customers in a cost effective manner.

      1. Jess Bachman

        If Groupon only highlighted businesses, it might work. But the users it attracted were total spend thrifts. I’m sure the Groupon discounts were much larger than the coupon books (which I sold in my youth ;). Groupon is lead gen.. the only problem is, they aren’t the glen gary leads.A strictly business-highlighting service would be which is pretty awesome, and run by some really smart people.

        1. LE

          The “problem” is that it sets people who don’t even care in the first place about discounts and gets them into that mind set. So it’s a larger pool of people than just the “discount or no business for you” set.Here is an example. There is a local sushi restaurant here that offers 30% off on sushi Monday and Tuesday night. (Very successful). However even though I don’t care about the 30% that I save I find myself choosing from the 30% items (and ignoring the full price items) because my brain is thinking that I am throwing money out.Eel sushi is $4.95 reg price and $3.50 with the discount. Do you think I really care about the price difference!!!! I actually don’t!!! But they have me conditioned to choose items that are discounted (I still get appetizers full price) and I have cognitive dissonance when I don’t choose from the discounted menu!!! And I am not even a person who typically cares about discounts!Likewise for the gas station here that offers 7cents off on Tuesday and Friday. I have almost gone for that but then I realize “so what are we talking about here anyway”. The fact that they offer that makes me feel bad (even though the money doesn’t matter we are talking what 20 gallons of gas!) for paying the regular price.I suspect that groupon has a same version of this going on.Groupon of course is also a “factor” in that I believe they advance money to cash strapped businesses last I read about them. That is a unique model that is more difficult for others to duplicate (requires a load of capital).

    2. PhilipSugar

      Actually I was not at all. You have an acquisition tool using an acquisition tool, so its like buying clothes at retail and then trying to sell them at retail as well.When you think about it Homejoy is really just an acquisition tool for its house cleaners.The hard part for them is what value do you provide so that the home cleaner does not go direct with the homeowner. That is always the hard part about an intermediate provider, it is easy for people to go direct and dis-intermediate the company.Then you probably had them having the cleaner also take a haircut with the homeowner.Btw: spent the last couple of days on the Strand at Charing Cross

  4. B. Llewellyn Shepard

    The way we get around this in my business – an agency, a service business – is to have a 90-day minimum contract length. Unfortunately we’re too much about retention. We throw so many resources at keeping clients happy that there’s often no time left over for sales… Growth instead comes from expanding existing clients.So our churn is effectively zero, but the business, being a service business, is in some real way limited by own mental head space.7 years in, if I had to do it over again? I’d absolutely have focused on a product instead of a service.

    1. JimHirshfield

      Even though I wouldn’t depend on locking clients in to contract term as a method of retention, 90 days is not a long contract term. So I’d venture to say it’s the quality of the service you’re providing which contributes to your client retention… And that’s a good thing, congrats.

  5. William Mougayar

    It’s like a leaky bucket. You can make it look full, but it’s artificial, and not sustainable. Growth without retention is not growth.A follow-on discussion could be around:- how to identify the reason for the churn? [e.g. customer interviews]- what happens when you “fix” what you thought was causing churn, and it’s still happening?- how do you realize when you need more “severe” iterations that might leave the small segment of users behind (the ones that aren’t churning), and try something different?These are tough questions to figure out, and I’m currently seeing them in a couple of cases I’m involved with.

    1. Jess Bachman

      And a very hard look at the acquisition cohorts, including messaging.If your acquisition channel is mis-representing the product, or attracting the wrong type of users… the bucket will leak out even if the holes are small.



    2. Richard

      Cox proportional hazard model is a great way to think about the issue.

      1. sigmaalgebra

        Okay:…Cox’s work is okay, but anything thatreferences Efron I’ll look at!First cut, I’d just model it as a Poissonprocess of new users arriving and another Poisson process of existingusers leaving.

    3. fs2117

      Exactly. It’s so important to peel back the layers. Too often, excitement around high-level vanity metrics cloud reality and real analysis falls by the waist side.Any CFO of a company with a recurring revenue model will tell you high level metrics, or even relative unit economic ratios like LTV:CAC, don’t paint a full picture. Component level analysis is incredibly important in order to capture the nuances across products, channels and vintages — e.g. a deep understanding of various customer cohorts, cost of individual acquisition channels (rCAC), payback period (GMPP), etc.You need to understand the ‘Why?’.

      1. TeddyBeingTeddy

        I always thought it was “way side”. “waist side” makes so much more sense! I learned something today…

        1. fs2117

          No, you’re right – it is “wayside”… chalk that one up to auto-correct on my dumb smart phone 🙂

          1. TeddyBeingTeddy

            The only thing better about Samsung than iPhone is iPhones stupid keypad and autocorrect errors. happens way too much

    4. TeddyBeingTeddy

      How do you know when it’s better to change the product, versus just changing the message (marketing,expectations, etc.)?

      1. William Mougayar

        It depends on the insights you gather. Without knowing your context, I would say if the product didn’t do what they wanted, it’s the product. If they didn’t know about something, it’s your communications. But it could be a 3rd thing: they don’t have the time to use it, because the ROI on their time or money isn’t there. You can ask them to hang on or retry once or twice again, really. Or maybe there’s a better product out there.

        1. TeddyBeingTeddy

          Good call, makes sense and thanks!



  6. Nathan Guo

    Fred, what metric would you look at for B2B sales?

  7. Mike Chan

    So true. Growth only comes after you can retain your early customers, or else it’s a leaky bucket. We’re facing the similar problem with my startup, ribl ( Fred, with this post and Friday’s article about the high-value niche use case, you’re reading my mind, it’s almost scary. Thanks for the great insight.

    1. William Mougayar

      what are your analytics telling you, so far?do you know where/when they are dropping off? have you talked to those that aren’t returning and those that are staying. any patterns?

      1. Mike Chan

        Typical chicken-or-the-egg problem for social apps – not enough content on the platform to be useful. We’re manually pumping in more and more good content but at this time it’s not scalable.And like Fred wrote, we don’t have that niche use case. We launched early to get the app out there to see if we can find trends of what kind of content bubbles to the top, but so far nothing stands out. So we may go back to the drawing board to focus on a very specific niche first.

        1. FAKE GRIMLOCK


          1. Mike Chan

            Hey Grimlock! You can argue social technologies like Snapchat, Facebook, and even Skype didn’t have high single-user value. But for the most part I agree.

    2. sigmaalgebra

      the high-value niche use case Yes, common advice is to “build something 100 people love”. But of course, in such a discussion, there are two quite different cases:(1) Hopefully the 100 are a good enough simple random sample of the huge market have in mind.(2) If the 100 are really in a “niche”, that is, typical of just a small market instead of a big one, then, the number 100 is nice but the business might have trouble growing to a big market.

  8. awaldstein

    Churn invariably comes down to brand value and in many instances it rolls down to pricing.When is the last time you raised prices to consolidate your base and reduce churn? (I learned this from the hands of the master, Umang Gupta himself.)

    1. William Mougayar

      but a tiny little startup has no brand value to start with, and might have a free product (for the consumer at least), so i think it’s back to solving the product/market fit conundrum.i think what you said applies once you’re up and running.

      1. LE

        Agree no brand value to start with.Also, in this case the “brand” is the particular service (and specifically people) that clean your home (or office) and it’s more like “can I rely on them to clean, at a fair price and not steal anything” (can I trust them). It’s kind of a stretch therefore to even use the word “brand” to describe this. Why? Because you only have one house to clean (most people anyway) and once you find the person who cleans you are set. This is quite different than deciding that you will go to the same hotel brand in different places that you travel or Starbucks or a clothing brand or a car brand. Kind of a stretch to put “brand” into this although I can see why it could help with nationwide marketing obviously.If you do a search for “house cleaning princeton nj” or “Orlando Florida” you find many possible cleaning services ranging from “Jane’s Cleaning Service” to “”. If I see “Janes” or “Patty’s” (that is also listed) why would I call actually? (I wouldn’t but that’s me). Jane seems like a real person that I can trust for my house.

        1. Jess Bachman

          Brand value starts the second you interact with a company/site. Just from the name “Jane’s Cleaning Service” I am already imagining them to be local, authentic, possibly honest. No need to wait until its a recognizable name. ..uhg.. they actually have to do more work to convince me they are local, authentic, and honest.

      2. awaldstein

        If you are touching a community, transaction or not, you are in brand building mode.Brand and pricing, brand and value start with the very kernal of market connection.And honestly, they change less than you think in the majority of cases.

    2. pointsnfigures

      Sometimes it takes customers a while to get familiar. They don’t care if it’s branded or not, just if it’s useful and they derive value from it. Getting to value is key. Most people will try anything once.

      1. awaldstein

        We are predictably on different sides of this one my friend.You look at marketing and brand building as something that you approach at a stage of growth.I look at it as part of the poise of building a market.WilliamM and I have never agreed on this. I think you are in his camp.More that one way to do this. The only thing that matters is if it works.Shoot me your email when you get a change on another topic.

  9. Guy Lepage

    Well I hate to say it, but probably one of the reasons Homejoy had retention issues is their brand was/is very weak. I feel that founders and VC’s truly underestimate the power of a strong brand. I’ve seen it in action. Night and day difference in culture.

    1. sigmaalgebra

      First suspicions confirmed: Homejoylooked like a fun story for the media buta bad business for everyone.

  10. TimGroot

    Would you be able to share some 90 day retention numbers that are considered good for social platforms?The higher the better of course but an indication of what the threshold is would be very much appreciated!

  11. OurielOhayon

    Great common sense. but it does not work with many type of services. Eg Airbnb where the cycle of order is longer than 90 days (unless you take a lot of holidays)

  12. David Barnes

    Quick metric for retention: DAU/MAU. (Daily Active Users divided by Monthly Active Users)What % of your total users turn up every day?If it’s >30% you’re on the way to building a daily habit for many.

  13. Jess Bachman

    Yep, 90 day retention is a good hard long look in the mirror.

  14. christopolis

    irony is that much of the non-retention is from giveaways funded by venture capital that is trying to buy a story

  15. Richard

    Yes, in a more technical spence, there is the adoption rates (new customers), growth rates (use frequency) and survival rates.I used home joy once (the lost $ on the transaction).They were clueless about how to increase my useage and at what point I had left the platform

  16. Carine Carmy

    Retention is critical, and perhaps more often overlooked because it’s a company problem, not just a marketing problem. For user acquisition, one person can spend more to make the numbers look great. Of course, optimizing the funnel will be a lot more sustainable, but you can still make traffic and sales grow with a leaky funnel (it’ll just be more expensive). For retention, you need product, marketing, operations, *everyone* working in sync to live up to the promise you make to customers. If competitors are executing better or setting a higher bar and delivering, short term retention tactics like discounts will be just that – short term. Retention work is harder, requires leadership, insights, and customer centric thinking. But it’s the only way to truly grow.

  17. Bernard Desarnauts

    Fred’s broader point on the need to really figure out product/market fit first – before stepping up on the gas is the most critical insight from today’s post. The 90-day retention / cohort is good for many but fundamentally the case that can be articulated for the vast majority of all new projects (& not just new startup) is the need to iterate/validate prior to full acceleration. Many talk about this Steve Blank “customer dev” process – very few take the time to do it.

  18. pointsnfigures

    Sometimes it takes longer than 90 days in certain businesses. Have seen it where a customer comes in, uses app, then walks away for a bit. When they return, they use it big time.

  19. Steven Kuyan

    Agree on all points. These two pictures are some of my favorite on the topic:-conversion/churn to show growth -impact of downgrades on overall subscriber numbersThe second one does a really good job showing how churn can be the killer of hockey stick growth.

  20. joewaltman

    Fred, could you please elaborate on what you mean by ‘acquisition cohort’. Is this based on when the user first signed up (i.e. users in their first or second month of tenure) or the channel from which they were acquired (i.e. paid search vs. referral).Thanks.

  21. daryn

    Additionally, with any marketplace, it is important to look at these retention rates on both sides of the table. They’ll affect each other, of course, and it’s risky to scale a sales force to recruit a bunch of businesses onto your platform if you’re not yet delivering on results and they’re all churning in 90 days.

  22. Michael Brill

    Unfortunately, for *early stage* companies who need more capital, focusing on retention can kill you pretty quickly. Dollars follow growth, not retention. That comes back to bite you, but you can rationalize to yourself and to your investors how you’re going to kill churn with the new ops platform your engineers are building. It’s much harder to convince them that your 30% growth will turn into 300% once the world sees how wonderful your service levels are. No entrepreneur likes to see customers bail on them… but the game is rigged that way.

  23. Pat Ryan

    Great point Fred. A lot of people say they are getting to the same place by focusing on net new user growth as the metric. The problem here is that it doesn’t unpack the metrics to an actionable level (retention vs new user acquisition) – and as noted below can be misleading because of aggressive acquisition filling a leaky bucket.The challenge in trying to solve for this is that many VCs push entrepreneurs on new user acquisition even when companies haven’t nailed product-market fit. By taking focus away from retention/NPS efforts they become less likely to solve for either.How should an entrepreneur handle a VC who pressures them on growth this way before they’ve nailed product-market fit?

  24. kevando

    RIP Turntable.

    1. Jess Bachman

      Yeah.. we’ve all moved on to Ello.

  25. TeddyBeingTeddy

    Brilliant post. thank you, sincerely.

  26. Michael Sher

    Fred. I assume 90-day retention for a service app like Homejoy is different than a productivity app like Centrallo? We track full, classic, rolling and return retention. We are also a freemium model so also what is the main retention metric we should be looking for?

  27. Salt Shaker

    The delta between Homejoy’s promo price on Groupon ($19) and their regular rate of $85 was likely way too rich. The low price likely attracted bottom feeders and not the kind of customer who regularly pays to get their home cleaned, hence low retention. I would hope they did a range of price elasticity testing w/ various promo offers to model conversion rates, but the Forbes piece doesn’t elaborate. It just paints the $19 offer as the kiss of death.”Protect the core,” particularly when it’s a helluva lot cheaper to retain a customer than it is to acquire a new one.

    1. JLM

      .The big problem with Groupon, and all of its spawns, is that it reprices the value proposition with existing customers.I bought ten Groupons to a hand wash car wash, where I had been going for decades, thus depriving them of my normal full cost revenue.Worse, now I wait for the specials and double down on them.Pricing for existing customers is a bad result of the desire to use pricing as a lure for new customers.I think Groupon can be highly destructive.JLMwww.themusingsofthebigredca…

      1. Salt Shaker

        Agree 100% JLM, but it also depends upon brand margins. Although on the surface a 10 pack discount seems a bit rich, diminishes the value prop, etc., perhaps the incremental volume generated more than offsets the reduction in price. Long term, clearly not a great brand building strategy, but car washes are gen a commodity biz, although certainly not for the fine care required of the BRC.I think Groupon is good for driving sampling/trial, but bad for driving long term brand equity, unless a company has a value positioning, operates w/ high margins, etc.My haircutter started his biz a year or so ago w/ a sick Groupon promo offer. Ridiculously rich. If you calculate rev to man hours in Year 1 he practically did 600 haircuts for free, yet he generated a 35% renewal rate. It would have taken him years to amass that kind of clientele w/ just word of mouth alone. Groupon can be effective or harmful depending upon where a company is in its growth cycle, whether it’s a “class vs. mass” brand, etc. Today’s mindset of generating short-term qtrly rev frankly feeds the beast! Groupon’s payout to its biz customers is additionally onerous (bad cash flow), while they get the benefit of “slipage” (paid for yet non-redeemed promos).

        1. JLM

          .I hear you and respect your thoughts but within your very words, I find all I need to indict the idea of using Groupon.The customer count may be the only redeeming element of the equation. Does it meet reasonable criteria of CTA v LTV when one takes into account the lost revenue from the existing base?There is a tendency to look for a magic bullet when marketing is often just drudgery and grinding it out.Inspiration — even silver bullet inspired marketing — is for amateurs while grinding it out is the breakfast of champions.JLMwww.themusingsofthebigredca…

      2. PhilipSugar

        Agree completely, and wrote about that when they first came out. In my mind many companies that use them are desperate.It is totally analogous to a payday loan. Its a really, really expensive loan, and the only reason you are taking it out is because you can’t afford to in the first place.

  28. Joy Kennelly

    I like your blog posts because they’re quick, easy reads packed with great information, kinda like a vitamin for startups.:) Inspiring me to blog more because it’s easier to swallow when it’s short and to the point. Thanks! Enjoying your blog very much.

  29. fdesroches

    What about the retention window for games? Are all games equal? Would you, say, put Halo in the same bucket as Trivia Crack when it comes to retention? I definitely agree that you must fix retention before you step on the gas, but figuring out what “good” retention looks like is the first challenge.

  30. sigmaalgebra

    It looks like Fred is saying that a startup should be well on the way to lots of users/customers, traction, revenue, etc. before trying to scale.Then I’m assuming that don’t try to raise equity funding until are ready to scale.Okay, I can see how this can work: Some startup is ready to scale, for some reason wants to grow very quickly, that is, faster than possible with just retained earnings, so needs some equity funding. Okay. I can see that there can be such businesses. Apparently Facebook was an example — lots of growth and equity funding before much revenue but now lots of users, revenue, and earnings.But, if the startup is ready to scale as described, then why take on the founder time and effort of equity funding, a C-corp, and a BoD? Instead, why not just grow more slowly, carefully, from retained earnings and keep the finances just dirt simple — 100% founder owned sole proprietorship?E.g., why not be like Plenty of Fish, good revenue and earnings right from the start, soon $10 million a year in revenue, from just one guy, the founder, two old Dell servers, and ads just from Google?Or, the US is just awash, coast to coast, villages to big cities, with sole proprietorships that make money and never got equity funding.High tech should just be an advantage, make money sooner, money enough to pay to scale. Besides, in all of business, software on a server farm is about the easiest way to scale; usually, just run more copies of the nearly the same software on more servers, say, if only at Amazon Web Services (AWS).E.g., the capex for a Web startup might need to cover just a server for as low as $2000, and that is much less than the capex for a plumber, electrician, grass mowing service, pizza shop, auto repair shop, auto body shop, etc. Looks like an advantage to me.Let’s be more clear about this, say, include some numbers, ones I find just astounding: Suppose we have a Web site and get some usage. Suppose we send a Web page for 400,000 bits; our upload data rate to the Internet is 35 million bits per second (Mbps); we have an average of five ads per page we send; we get paid $2 per 1000 ads we send; and on average 24 x 7 we half fill the upload data rate. Then we send on average 24 x 735 * 10**6 / ( 2 * 400,000 ) = 43.750pages a second and5 * 35 * 10**6 / ( 2 * 400,000 ) = 218.750ads a second, get paid2 * 5 * 35 * 10**6 / ( 2 * 400,000 * 1000 ) = 0.437,500dollars a second, get paid2 * 5 * 35 * 10**6 * 3600 * 24 * 30 / ( 2 * 400,000 * 1000 ) = 1,134,000dollars a month.For the35 * 10**6 / ( 2 * 400,000 ) = 43.750pages a second, that takes, what, just a few servers at ballpark $2000 a server.Suppose to start can send 1/10 that many pages a second from just one server. So, that would be1,134,000 / 10 = 113,400a month from one server. So, right away buy the other 9 servers with money enough left over for more than one nice dinner at a local bistro.The 10 servers? If don’t want to use AWS, then put them in a spare bedroom, on a wire rack shelf unit for $100, with a big A/C in the window — have an electrician install a 200 A circuit breaker box in the bedroom.After 90 days at$1,134,000a month, with one guy, who mostly eats at home or at McDonald’s, what’s this about a Series A? Why bother? Or closing the Series A might take another 90 days, and by then the guy has $6 million in pre-tax cash in the bank. What’s he want a Series A for?I’m serious: Millions of successful sole proprietors — coast to coast, villages to big cities — plan their startups to get earnings early on; why not technology startups? I’m not getting it.

  31. Vasu Vadlamudi

    Totally true. As Zynga matured the #1 thing new game launches cared about was NOT install growth – we knew the installs would come if we had built for the correct audience (at least before per-user acquisition costs in a maturing ecosystem became a huge constraint). What we cared about most were early retention and engagement, because in a daily-use app, those metrics strongly predicted long-term retention and the maximum userbase “ceiling” of the product.For early retention, if your product can be used by a normal user on a daily basis, the “40/20/10″ bar is a solid goal. (Of 100 installs on D0, 40 return on D1, 20 on D7, and 10 on D28.) It’s not clear to me that similar norms have emerged for apps with weekly or monthly usage patterns like Homejoy, but would love to hear about them.The other interesting thing about these metrics is that they are a blend of platform/”grind-it-out” optimizations (notifications, retention emails, social flows) and pure product design. No matter how good your optimizations, if your product design is garbage you will not hit and sustain really impressive retention numbers… which is why I get nervous when a startup believes they can solve everything with growth hackers/retentioneers. I say this as a product guy.

  32. Neima

    Totally agree! Retention is much harder than acquiring new users. You actually have to spend time and understand why they want to leave, do your best to retain them, and then also implement the fixes to the product. Its hurtful to the company’s ego to think that they can be wrong…But, since a startup isn’t meant to be easy, retention, especially of early customers, can give a startup great insight on how to improve and build a really great product.

  33. Sriram Yadavalli

    In markets like babysitter/cleaning (& any other service where there is possibility of differentiated offering), there will be disintermediation. IMHO, this was a big factor in HomeJoy’s decline

  34. Douglas Crets

    I am probably not on people’s good sides for the amount of time I will rail against growth hacking, but this is my primary grievance with growth hacking: it slices off retention because it focuses too much on metric gains, which I think are only there because they are tied to investor money.What growth hacking ignores — probably because it’s main body of practitioners is Silicon Valley VC-focused / incentivized, and (sorry), young — is that marketing is nearly always a long term narrative and it is directly tied, in digital marketing, to customer service, retention, and ongoing connection to users who have been around a while.Regis McKenna talked about this a lot in 1991, when he wrote “Marketing is Everything.” I think that growth marketing, or growth hacking, is the misalignment of engineering mindset with search for money. Marketing should always be about aligning customer values and feelings with product attributes and service.

  35. mikenolan99

    The founder of a great start up told me about the concept of “regrettable attrition.” We were speaking of employee turnover… But the concept applies here. Keeping the right users is more important than total number.

  36. Shanti Bergel

    Retention is extremely important of course. But, I don’t care much for the idea of looking at any metric in isolation. There is no “one number” in my opinion, although you might be able to get away with three – retention, monetization, cost of acquisition. All of which can function differently at scale.Great retention stats may get you a second funding meeting but they won’t build your business without a corresponding monetization vector that enables profitable LTV>CPI arbitrage…except in very particular cost of acquisition circumstances. Zynga masterfully exploited the early days of the open Facebook API to amass hundreds of millions of free installs. When traffic is free or discounted, the bar for retention and monetization is correspondingly lowered. Those “free growth” circumstances are often temporary or scale capped however so, in the end, it pays to understand both the cost and the revenue side intimately.



  38. Jon Katzur

    Another added reason to care about retention over growth is that the best products see significant user growth from word of mouth growth of their current (happy customers/users). This is especially true, actually, in enterprise tech, where you can “buy” users with extreme advertising and sales spend, but without a product customers need and want to talk about as an engine for growth you will be in trouble.

  39. JonathanAquino

    Doesn’t this article assume that the product is web-based? It wouldn’t apply to a startup like Tesla, for example. It is limiting to assume that startup == website/app.

  40. andrewdever

    Growth & Retention are one of the same: shared experience loops that leads one user to bring another in AND keeps them coming back.

  41. Girish Mehta

    Burst out laughing !!! Ah…what wit.