Winback Campaigns vs New Customer Acquisition

One thing I’ve been pushing hard in my conversations with entrepreneurs, CEOs, and management teams I work with is the efficacy and capital efficiency of winback campaigns vs new customer acquisition campaigns.

Every company wants to grow their user base and increase retention.

As mobile becomes a more difficult environment to grow in (maturing market, more competition, growing dominance of the leaders), we see companies spending more and more money on new customer acquisition. While that is necessary, it is not likely the most capital efficient way to grow. Winning back churned out users can be a lot more cost effective if done right.

I like to ask these questions:

  • how many downloads have we had on our app in total?
  • how many of those users who downloaded our app are still active every month?
  • how many of those apps are still on a user’s phone?
  • can we notify those users?
  • can we email those users?

Generally speaking, an app that has been in the market for 3-5 years will have an active user base that is 10-30% of the total downloads. There are a few exceptions where that number is a lot higher, but those are exceptions (and great apps).

So that means that 70-90% of the people who have downloaded your app are not using it.

Some of them will still have the app on their phone and can be notified.

Most of them will be reachable via email.

So that is the winback opportunity.

You can’t just spam those users with notifications or emails that say “hey, try out our app again.” That won’t work and will lead to users quickly uninstalling your app if they haven’t already done that.

But you can look for targeted opportunities to tell these churned users something useful to them and invite them back to get more useful stuff.

This could be an offer for a big discount if you have an e-commerce related app. This could be a geotargeted offer/insight if you have an app that is location oriented. This could be the fact that a close friend is active on the app and an invite to join them.

These winback campaigns, if done properly, can be incredibly effective.

If you aren’t doing winback campaigns, you should be.

If you are doing them, think about how you can do them better and do them more frequently.

This is about the lowest hanging fruit out there in the mobile app growth landscape.

#mobile

Comments (Archived):

  1. awaldstein

    Companies are either fighting to gain market share or fighting not to loose it.I attribute this smart phrase to Albert btw and it covers the continuum of activities of what is smartest at any point in the lifecycle of a brand.Been talking this one up to my clients as well lately.

  2. Typo Alert

    “As mobile becomes *are* more difficult environment”should be”As mobile becomes *a* more difficult environment”

    1. creative group

      Sorry to inform you but the responses reflect the active participants on this blog are literate and realized the typo.Thanks for your input.

    2. fredwilson

      thanks!

  3. William Mougayar

    But you also need to find out why they churned via direct interviews with a sampled variety of churned cohorts you have. This informs you to devise better “winback” campaigns.And the big assumption here is that your product has changed or evolved, so you can have something new to talk about. Otherwise, you’re just annoying them with – “hey try us again, in case your needs have changed.” You can give me 50% off on bad coffee, and I still won’t drink it.And maybe a special offer is not enough. Maybe it’s a better articulation of the value or benefit of using that App, and you need to further sharpen your message.So, before these winback campaigns, sharpen your product, and sharpen your message.

    1. pointsnfigures

      Exactly what I was thinking. Take the opportunity to find out why they churned. If they can’t tell you, hopefully the entrepreneur is good enough at pattern recognition to piece together a cohesive reason why after interacting with a number of churned customers-because as you know sometimes customers have a very hard time articulating exactly why they don’t like your product.

    2. creative group

      WM:there is no need for us to reiterate what you have excellently illuminated. The service or product needs to be useful and evolve to maintain users.The question is why are the users dropping an app that was once useful enough to download. No assumptions required just ask the churned users.

    3. Justin Raymond

      Ahhh yes…the ol’ Winback / Re-activate / Resurrection campaigns’ … from my experience, deep cohort creation prior to outreach is crucially important – and, every business should have different cohorts based on use, frequency, time of use, location of use, etc. You only have one chance (maybe two) to effectively re-engage with the consumer so take the time to ensure you are maximizing relevancy of message. Timing of the outreach is also crucial…what is the most relevant day of the week/month for your service? Does weather impact relevancy? Lead up to events / holidays etc could play a big part in your ability to successfully resurrect the user.

      1. William Mougayar

        spoken like a true warrior 🙂

    4. Twain Twain

      You and @pointsnfigures beat me to it.I was going to comment: I always ask, “Why?” as my first question. The how many and what flows from this.[Maybe this has to do with being a Maths and b-sch grad. I know there are a whole bunch of tools that can help work out the “How many” and “What” questions.The “Why?” is a harder problem to find the tools to solve.]A lot of the time entrepreneurs focus on torquing the quant metrics (increase downloads, increase engagement etc) but not enough on the qualitative metrics like “Why would someone sign up and use this over and over again? Why are they feeling closer to us rather than to a competitor? Why would some bigco want to be our strategic partner?”In some ways, I see “How many” and “What” answers to be about system efficiency (churn rates) but “Why” answers about user effectiveness (loyalties and appreciation of the brand).Utility (usefulness) = functional matrix of (efficiency x effectiveness)

    5. sigmaalgebra

      > And the big assumption here is that your product has changed or evolved, so you can have something new to talk about.Good. Early in my career, I was developing, documenting, and supporting software, and finally some of the sales staff explained that one of the main uses for product enhancements was just an excuse to return to customers and to tell them about the new things and, then, of course otherwise sell them as usual.Lesson: Always be (standard first two words for a lesson in business!) adding to the product.

      1. LE

        Very true. Somewhat related is the concept that when a customer has problems with your product or service it can actually be good rather than bad. I have personally found this to be the case. Simply because the customer then rates you on how quickly you solve the problem and how attentive you are as well as the solution. If there are never any problems, you tend to be forgotten and don’t have much of a chance to earn “brownie points”.Each of us probably sees this frequently in every day life in big and small ways.I am reminded when someone who was working for me made a mistake that screwed up an important piece of software. But then they worked very hard (and even delayed a trip by taking a later train) to get it fixed in time for a particular deadline. In another case a consultant was leaving and traveling over the summer but informed me in advance that he would be available even if he was in different time zones and that if he didn’t get back to me immediately (as he normally did) not to worry. Each of these are examples of turning what could be a negative into a positive.

        1. sigmaalgebra

          > Somewhat related is the concept that when a customer has problems with your product or service it can actually be good rather than bad.Yup, I have a friend who noticed that: He saw a little company that had a mold, some heat lamps, and a vacuum pump. So, put a sheet of plastic on top of the mold, turn on the lamps, soften the plastic, turn on the vacuum pump, “giant sucking sound,” and, bingo, presto, the hull of a small boat. Add seats, hardware, etc. and have something for sale.So, he got to be a distributor. To sell, he went to people selling other things, maybe outboard motors, RVs, cars, etc. and had them take on selling these little boats, nice profit margin, as a sideline.So, when filling the first order, he made sure that a part or two was missing from the boat he shipped. When he got the irate phone call, he went into his well planned program to respond, e.g., jump around like spit on a hot griddle, to “make it right”.It was all a deliberate act, but he thought that it served a purpose.

          1. LE

            So, when filling the first order, he made sure that a part or two was missing from the boat he shipped. When he got the irate phone call, he went into his well planned program to respond, e.g., jump around like spit on a hot griddle, to “make it right”.Well if you have ever run a company which is staffed by ordinary city people phoning it in and making a custom product (as opposed to the type of Ivy League high achieving graduates working at startups) you will find that there is no need to “make sure a part or two was missing” because they inevitably fuck up with predictable precision just doing their every day jobs!!! I kid you not! Like clockwork. We even had a saying for it. It was “Wheel. Of. Fuckups.”. Would To the tune of the game show “Wheel of Fortune” intro. You could bank on it.

          2. sigmaalgebra

            Yup, that’s been my guess at how Microsoft got so many security holes in their software that, for some reason, they believed, say, providing updates to fix the holes, they liked: Just don’t work very hard in the original design and coding to be sure don’t have any such holes. Then, have lots of holes. Don’t know where they are just yet, but let the hackers between the Black Sea and the Baltic Sea find and exploit them and then the users to suffer from them and like Microsoft for the patches. Hmm ….And, this way, no Microsoft employee can go public with evidence that the holes were deliberate.Then when have nearly all the holes patched, end support, come out with a new version, tell all the customers that they should upgrade to the new version because, as errors were patched in the new version but not the old, hackers would suspect that often the same hole was still in the old version and would exploit that. Hmm.Some evidence: As of a year or so ago, Microsoft was still fixing “buffer overflow” errors. Gads, that’s a Programming 101 error. That such an error got through at all makes some of the Microsoft design, development, desk checking, and testing work look like a WoF, at the level of Gates, deliberately so.This suspicion got to be so common that the Bond movie Tomorrow Never Dies explained with glee that a new software offering was “full of bugs, forcing users to keep upgrading for years”.That’s one way to have continuing customer contact.Maybe one day early at Microsoft some mid level executive ran into Gates’s office screaming that some Microsoft software was full of security holes and Gates quickly saw that the situation, encountered by accident just from pushing code out the door ASAP, was not a problem but an advantage as in “it’s not a bug, it’s a feature.”.

          3. LE

            Back in the day was typical for copiers to constantly break down so that you would have to pay for a service contract to keep them from braking down. I am sure you remember the “copier repairman”. Once distribution changed to big box stores for copiers all of that changed with the exception of copiers that still were large enough to be sold direct by salesman. But even those copiers got more reliable because they had to be more reliable. (Ditto for cars once manufacturers had to offer bumper to bumper warranties because their competitors did, no more corner repair shop except in ghettos and poor neighborhoods typically).As far as Microsoft Occam’s razor is closely tied between “they suck” and “they are being opportunistic”. Differential diagnosis so to speak.The short amount of time that I worked in Silicon Valley for a tech company with a product, I can vouch for the fact that they knew the product sucked but it was still released knowing that the sheep that bought it would just chalk it off to typical computer guy stuff they didn’t understand and a battle they didn’t have time to fight. This all started with PC’s by the way. Back prior to that shit was done right because of the profit built in.

    6. Douglas Crets

      I’m in agreement with William on this. Interviewing, research, feedback, and identity research.

  4. LIAD

    Linguistics matter.I like the phrase ‘winback’ rather than the more prevalent ‘reactivate’. Never heard it before.Winback – implies you’re dealing with a living breathing animate object and have to work and deliver value to regain their attention, mindshare and usage.Reactivate implies you’re dealing with a robot and are just looking to press the right button so you can extract rather than provide value.Winback. I like it.

  5. William Mougayar

    I would add another dimension to the Winback strategy: You need to also Win back your Brand in the marketplace, and that will get you incredible market pull, if done well.Some startups are focusing too much on product and digital marketing, and forgetting the corporate/brand marketing aspect, because they know nothing about that. They are falling in what I call “the digital native marketing trap”. They think it’s all digital, and bottoms-up marketing, and they forget about top down brand marketing.You can push push push, but at some point in time, you need to grow your brand, and that will pull customers and users to your product with less effort into it.

  6. jeffyablon

    There’s a very basic level at which the statement “it’s easier to maintain a customer than it is to find new ones” is true. Not a little true. Not cliche-level true. TRUTH. FACT. (And data analysis aside I’m a big proponent of that idea there are no—or at least very few—facts).And you’ve missed something crucial here.Once upon a time, I did a few years at Verizon. One of my gigs there was as a Business Development Manager in their WinCities programs, trying to effect Winbacks (it’s actually the word we used).That sounds like a sales job, and of course ultimately that’s exactly what it was. But we had only soft quotas; the job was community re-education about coming home to Ma Bell.We had a serious leg up in that we’d never really lost the customers; we were merely trying to work with them without the middlemen.This plan failed. Unequivocally. Which makes no sense until you realize that there’s a flaw in the pitch.Assuming they haven’t gone out of business or left your usability footprint (internet corollary: leaving a country that has the rights to broadcast something for one that does not), when people leave you, assuming it wasn’t because you chased them away with indifferent service or high prices, it’s either because they’ve realized they don’t need what you offer or have found a better way to “get it”.This makes me think, among others, of Foursquare.Creepy spy-on-you issues notwithstanding (and we all just accept that now anyway; hello, Google!) Foursquare was cool. Real thoughts about real places you had really visited from real people. Wow!But once you leave them, you’re gone. There’s no way Foursquare can win back prior customers, because they no longer offer anything unique enough to make using them matter to you and you probably realize EXACTLY that, regardless of the reason you had when you left them.This applies pretty much across the board. You haven’t offered something legitimately NEW and also found a way to make me realize that’s the case? Then I’m not listening.That’s the fallacy in your point here, Fred. Yes, people with whom you have a connection are lower hanging fruit than strangers. But that fruit is … bruised. You may be able to pick it, or at least scrape a bit of it into a receptacle and make compote, but unless there’s something GREAT about that fruit (and you’ve added something to your soil to nurse its tree to better health), you’re better off planting new trees.Of course, those trees can be hybrids and the first people you try to sell them to might well be people who’ve bought from you before.The point is this: Winbacks don’t really work. What works is finding a way to make those people with whom you have a glimmer of a relationship realize you’ve got something BETTER than you used to … not try to “get them back on board”.

    1. PhilipSugar

      I think you have a really good point using Fred’s analogy of winbacks are low hanging fruit: You have to realize the fruit is bruised.Now it depends on why they left: apathy? or for a competitor. If it is the latter you see companies like Verizon that are trying winbacks for T-Mobile defectors. But you better have a better product. (and Verizon is improving their global roaming, but still not great)That is the beauty of competition.

      1. LE

        You have to realize the fruit is bruised.Exactly. Standards become higher. Fruit is more likely to be trigger happy and think “same old shit”.

      2. jeffyablon

        All well put, Philip. But let me go even further in defining the terms of my story: I didn’t work for Verizon Wireless, but for Verizon … the real Verizon, if now the smaller one. The wireline company.As an incumbent post-Bell, when someone has wired “phone” service (or any service that uses the phone lines or the phone poles or conduits), they have service from Verizon—as opposed to the way things are in cell phones, where there’s still a tie in to that very system, but the towers you get access through belong to the company you buy service from (err … until they don’t, but T-mobile uses different towers than Verizon Wireless).So Winback in Verizon’s case was a disingenuous concept from the beginning. “Cut out the middleman, have greater control, and make more profit”, but the customers had never stopped being our customers.So that’s even LOWER hanging fruit, and the bruises, frankly, were always actual bruises. I live in NYC where both wireless carriers have a strong signal, and even here switching from VZW to T-M isn’t ever about service quality, but simply about price. That situation is a simple winback; make the product cheaper and back the customers come.But price at least is a meaningful change, right? My point is that this conversation as Fred started it seemed to be about change through existing-relationship marketing without any real change to make it meaningful (the model I was talking about at VZ). And that’s just a mistake.

  7. Scott Barnett

    It costs 5-10X more to acquire a *new* customer than it does to upsell an existing customer. Yet most emerging businesses put more effort/focus on acquisition. It’s psychologically sexier and easier (a bigger perceived pool of potential customers) but academically, going after folks that already know you is much more effective, and they are more prone to tell you if something is going wrong. You can also (as William) mentioned go “deeper” with existing customers to get to root cause of churn.High touch wins with emerging businesses.

    1. Richard

      Missed opportunities for all those data scientists !

  8. Conrad Ross Schulman

    Happy Sunday everyone!

  9. OurielOhayon

    question #3 and #4 can t be answered correctly on iOS. it is impossible (even less on ios9) to know if an inactive user has the app on the phone and it is impossible to know if a user has revoked push tokens (apple does not provide that information). It is also not possible to know where on the phone your app is placed (homescreen, later, in folder)….Emailing users who would if there is registration and end email has been collected. by experience re-activating lost users is very very hard and sometimes better just getting new ones.

  10. JimHirshfield

    There is also the phenomenon of early product releases not being fully baked enough for users that aren’t early adopters. These users try, and don’t like (yet?). A more mature product is what they want. There was an article recently on this topic; the premise being that you don’t want to be featured in the app store too early in your product’s life… it leads to too many users abandoning the app due to unmet expectations.

  11. jason wright

    ‘churn’ means nothing to the individual user, and all users are individuals.what word(s) would the user use to describe their behaviour? that’s the insight to begin with when working on a ‘winback’ strategy.

  12. sigmaalgebra

    If you can contact those former users at all, then, of course, invite them into your Round Table of Valued User Advisors or some such and have them complete a questionnaire that can tell you at least some of why they are no longer using your product.

  13. Richard

    Winback is the operative term. Even in cases where I’ve spent 1000s of dollars (instacart), I’ve heard nothing. This is when I know “kids” are running the show.

    1. awaldstein

      Not a believer that they are going to win for many reasons–that and the fact that what they provide is really not a challenge for the stores themselves to own and provide at a much much smaller premium to the customer.

  14. reece

    hehe i can identify at least two companies in the USV portfolio doing this (recently)…more important than anything is that the company needs to provide value to the customer/user… just saying “come back” or “please buy more” doesn’t really workthe apps/products/subscriptions i stay on, or that “win me back,” are the ones that find a way to provide great value to me (this is where network effects work best. kudos to USV thesis. ex: i’d sort of written off Facebook for a long time, but the network pull is too strong and they recently won more of my attention. the key product move was keeping Messenger tied to Facebook itself on desktop web [unlike on mobile where it’s its own app], so that when i use Messenger on desktop, i was forced to see new iterations of my Facebook news feed, which had improved)on the flip side, my experience with Netflix was poor. in the early days (we’re talking DVD’s mailed to my house here), i gave Netflix a TON of user intention. had a list of 100+ movies that i wanted to watch on their service. eventually, i just kind of stopped using it, even as streaming became more mainstream. Netflix was happy to take my money, especially while i wasn’t actually costing them anything (no shipping, no bandwidth). i think that went on for over a year! Netflix had all this user data… why not act on it and remind of the movies i said i wanted to watch? eventually i just said “fuck it” and cancelled and they haven’t won me back sinceanyway, if any co’s are listening – don’t just try to ‘win back the lowest hanging fruit’… go all the way*** and show the customer why they paid attention to your product in the first place*** https://www.youtube.com/wat

    1. JLM

      .Wow!That is an inspiring bit of video. I hate the idea that I’ve been alive for as long as I have and have never seen that video.That video should be required viewing for any entrepreneur. Every entrepreneur.Stronger than a hundred acres of garlic!Well played.JLMwww.themusingsofthebigredca…

      1. Matt A. Myers

        Wow. You are right. Great poem, great video.

      2. Matt A. Myers

        Goosebumps.

    2. LE

      why not act on it and remind of the movies i said i wanted to watch? eventually i just said “fuck it” and cancelled and they haven’t won me back sinceWhy? Because you are dealing with a fast growing company (rising tide floats all boats) where most likely it’s totally “shiny ball” and not only do people move on in jobs at that company, nobody gets any pat on the back for solving any particular problem if it’s not their job.I actually have this concept that I want someone to do (not going to be me) which is called “Tell CEO”. It is for all of those little suggestions that customers are aware of that they want to make sure that the CEO of the company knows about. Because surely they figure if the CEO knew about some of the things that are happening at their company, they would consider taking the steps to deploy resources to solve the problem.The “corporate” way that CEO’s currently deal with some of this is by stupid fucking popup annoying surveys and/or phone surveys done by someone in a cubicle reading a script. Typical lame corporate shit that turns off the most valuable source of information.

    3. Twain Twain

      Half-baked,Half-hearted,Half a step untaken,Are the half measures,Of those who Have Not,What the Haves do.(C) Twain, 23 Aug 2015.

  15. JLM

    .The real issue here is the creation of “loyal” customers — the 10-30% that Fred identifies as those who actually use and, hopefully, like your product.While I agree completely with Fred and the many commenters who have discussed the efficacy and importance of winning back lost “tryers” — which is what many of them are, there is an earlier point in time in which the same effort can produce better results.The objective of all marketing efforts is to create “tryers”. TRYERS! OK, then what?The next step is to transform tryers into “loyal” customers and therein is the rub.What is a loyal customer?They are that 10-30% who will behave like this:1. Loyal customers will use your product v all competitive products.2. LCs will recommend your product to their friends, strangers, visiting space men. Rising to brand evangelism.3. LCs will invest time and energy in the relationship — that’s the key, there has to BE a relationship.4. LCs will pay a small premium for their favorite product.5. LCs will “forgive” you when you screw up. Have a bug? They will not dump you. When you fix it, they will love you even more because you listened to their concerns.6. LCs will tell you the truth.[Manager of a restaurant comes up and asks, “How was everything, folks?” You say, “Great!” On the way out, your dining companion says, “Why did you say that? It sucked. Why didn’t you tell him?” “Cause I didn’t want to screw with it. I’m never coming back anyway.” Then, of course, you castrate them on Yelp.]7. LCs will engage with you.Before you go chasing folks who have thrown you to the curb, understand how many LCs you have and why? Focus group the Hell out of this and get to know exactly why they love you.THAT IS THE SUCCESSFUL BEHAVIOR YOU WANT TO REPLICATEModel that successful behavior — which your LCs are telling you works — before you go chasing the folks who say you suck.When you do go chasing the folks who have spurned you, replicate the behavior that won you the loyal LCs.Learn why you are successful before you try to learn why you failed. Model the successful behavior with the winbacks. Know the solution before you confront the problem.JLMwww.themusingsofthebigredca…

    1. LE

      Manager of a restaurant comes up and asks, “How was everything, folks?”My wife has learned to accept it now but when we were first married she was totally bothered by how much I complained about things. (Of course I was typically paying and it wasn’t her money but that’s another story). For example my drink is a long island ice tea. If it’s well made I can only drink 1/3 of it and then I have to stop. If it’s not well made I can drink the entire drink because it’s weak (and I won’t even feel any effects either). So since I know that drink so well I tend to make comments about it (as it can cost anywhere from $8 to $16 depending on the place). In one restaurant the manager had come over a number of times and apologized and he would say “well I watched the bar tender making your drink and they definitely poured in the right amount of alcohol”. To which I would answer “then your alcohol is diluted take my word for it”. She didn’t like the conflict so I had to stop complaining. Then one day there was an article about restaurants in our state being fined for diluting drinks which more or less proved me right. And all the sudden after that the drinks in that restaurant improved greatly and the problem pretty much went away.Now my wife isn’t bothered as much and actually complains herself. Not because of the “fines” but because over time she has become desensitized and has realized that she also does not want to get pushed around and just “be quiet” when there is a problem. So we always speak up. Not to get a price reduction but just not to get pushed around. Who cares what the waiter thinks? We are always nice about it by the way.

      1. JLM

        .We may be related. I always complain if I’m not happy but I always ask for the manager and never complain to the waiter.I also do the same thing when I have great service.My wife goes to the car and lets me do my thing.Because I used to own a lot of downtown ground floor locations, I got to know a lot of restaurateurs. They want and appreciate feedback in person.Yelp has changed how this works and reputation management is a big thing.Austin is a good foodie town and every day there is a new restaurant opening. Last night at dinner with some friends we must have spent an hour jawboning about new restaurants.JLMwww.themusingsofthebigredca…

        1. LE

          My “advance team” [1] who I drop off first (so I can park the car) know that only certain tables are acceptable to me. If they are not seated at what I consider an acceptable table when I walk in, I will walk right over to the hostess and tell them I want a new table and everybody will have to get up and move.Another thing that I will do is book a larger table (say for four people) when I only have two people. Then when I get to the restaurant I might say that “only two people are here” so I have a better chance of getting a four top (or a booth) instead of being seated next to people I don’t want to be next to. This doesn’t work if the restaurant is super busy though in that case you can end up with no table because they won’t easily give up the four top. So you have to shelve the strategy.What is interesting about what both you and I do is really the inverse of the “party in your brain”. That is neither of us really care what others think. Whereas your wife, my wife and my mother, are totally bothered by the “evil eye” of a complete stranger even putting that stranger above a person that they love!![1] My wife and/or my daughters

    2. creative group

      JLM:had to reread. Great observations. Notes any business can use to become better and retain patrons, users, etc.

  16. LE

    You can’t just spam those users with notifications or emails that say “hey, try out our app again.”USV of course has an upside in the spam fight (with http://returnpath.com/solut… and http://returnpath.com/solut… ) however the truth is that the major mailbox provider (in particular google) solely get to determine what gets delivered and what does not.It has gotten to the point where mail gets blocked quite arbitrarily and there is no court of appeal. So for example a situation where I simply send my wife (who I communicate regularly with obviously) an email with just a link (“read this story”) might go to her spam box on gmail. Totally fucking ridiculous. Or in other cases legitimate mail from low volume senders with established domains gets blocked. And in cases where you have a new IP for a server you have an even bigger problem. It’s shoot first, ask questions later. All of this of course benefits google and gmail because it’s not Larry Page and Sergey Brin’s “don’t do evil” problem it’s the person running the server and/or their customers.

    1. Matt Zagaja

      In general if an e-mail address is in your Google address book it should not end up in spam. Really strange that it is. Sometimes I’ve had to manually create filters to keep stuff out of spam (recently with Amazon which apparently isn’t meeting some kind of authentication standard and thus getting spam boxed). The decisions that the machine learning algorithms make aren’t arbitrary but based on their algorithms and data (see http://gmailblog.blogspot.c…. Sadly it is not yet perfect.

      1. LE

        Sadly it is not yet perfect.Sure I guess Google doesn’t have enough money or enough people working there to take on this issue and not have it happen. But it does happen that is for sure.Likewise Thunderbird email client also identifies things as “scams” that are not even close to being scams. I guess they are to busy doing other shit as well.

  17. Semil Shah

    I remember talking about this with you at our lunch earlier this year. This is the type of thing that hopefully gets thrashed around in a BoD meeting and then monthly through updates, but what’s sad about today’s environment (at least in seed) is that this data is rarely if ever shared, so the conversations often don’t happen.

  18. Matt A. Myers

    Brewster recently sent me multiple notification emails within 2 days… I never gave Brewster my email – it was other people of course who they got it from.I don’t think I ever subscribed either – so I’m not sure how they’re safely sending these emails. The newer Canadian email regulations are pretty strict.

  19. Pete Griffiths

    Great post – thanks.

  20. Gregory Magarshak

    Yes! There is something to be said about this from the point of view of system architecture.If your app has “smooth signup”, whereby you invite users via SMS first, give them a unique link to a web experience, and only *then* give them an option to download the app, then you will convert your viral loop better. There are several reasons for this:* The main reason is that it’s better to confirm the user’s mobile phone as soon as they click a link. Instead of taking them to an app store with a picture and description, where many of them will exit the flow. That way you have multiple chances to get them back, by sending them transactional notifications when something happens involving them.* Because by confirming the user’s account in ONE step, you can plunge them right away into a social web-based experience where they have friends. The reason to download an app should be organic, e.g. to record a voice response to a conversation, to let the app access their own address book, etc.* It’s actually more secure and doesn’t even require passwords. When the user finally installs the app, you can use OAuth with your web-based cookie to recognize the user and make a new infinite session for them in the app.I mean this is like 1 of literally 50 different things that would make the difference between an app that is consistently viral and engaging, and one that’s relying on luck to make it.PS: I asked a question on Quora recently about the legal aspects of the SMS messages and opt-out autosubscriptions ( http://www.quora.com/Under-… ) … it seems to be totally legal if the platform sending the messages is owned by a nonprofit. But also it may be legal if the SMS messages are transactional in nature and originated by the user’s friends — the argument would be made that we could have given the user a code to forward to their friend but our primary purpose in sending the message ourselves is to make the process much more secure. Any lawyers here have some advice about this?

  21. Douglas Crets

    There was discussion for a very long time and it fades in and out of the importance of storytelling, but I believe that storytelling and marketing narrative are really how you grow a company in the long term, and increase retention. I have written in two ways about it. First, on Pando about how real marketing needs to be more like journalism: https://pando.com/2013/09/2…And recently on medium about how I am trying to develop a practice called Receptive Marketing: https://medium.com/@douglas… /what-is-receptive-marketing-4365e795d7cbandhttps://medium.com/@douglas…

  22. Vasu Prathipati

    @fredwilson:disqus through your portfolio or the great vine, do you know of any vendors that can help marketers design more effective win-back programs?You mentioned compelling, relevant content (i.e. new feature releases) but is there anything else that goes into a quality win-back program?

  23. Michael Elling

    Getting, keeping and stimulating demand is a bitch!Every additional step in a process cuts your demand in half. 3 extra steps = 1/8th of the original addressable market.If you haven’t built feedback and viral loops in your product you’ve lost.If you aren’t analyzing usage and knowing why customers have left, you’ve lost.Instead of viewing it as a “winback”, it’s better to figure out what went wrong or what you didn’t plan for or execute properly to begin with than to try to continue selling a sub-par experience.

  24. Pankaj Garg

    Google now is one such solution..it is going to do exactly this in future.

    1. Mike Geer (MG)

      How best to use Google Now for re-engagement?

      1. Pankaj Garg

        Sorry Mike.. I am suggesting that Google Now has the potential to be a tool to do the same.

  25. Mike Geer (MG)

    Thanks, Fred! The concept is definitely sound and something my team has been starting to look at more and more. We’ve had 350 million downloads of our app Hotspot Shield and our all time retention numbers are in line with your range.Can any of group here offer any specific examples of messaging and channel that worked well for them and even specific conversion numbers would be awesome. We have recently changed our iOS app from in-app paid after a 7 day trial to fully freemium (ad supported after the 7 day trial), so we’ll be trying out some push notifications and emails around that theme. Happy to come back and share the results with everyone here. Of course every product is different and definitely agree that understanding the cause of the churn and also segmentation of the “win back” messaging is important, but some more specific examples would be great for sparking some brainstorming.

  26. Marissa_NYx

    Timeless advice! One of the most practical strategies most if not all Internet companies can adopt. Win back was a core part of the BD culture at one of the telcos I worked with, the board treated with the highest priority. So should the boards of modem Internet companies. It should be a permanent item on the board agenda.

  27. CathleenRitt

    Janet Choi (@lethargarian on Twitter) from Customer.io recently wrote about how to use email to win back leads, http://customer.io/blog/win

    1. Janet Choi

      Thanks Cathleen! Great points made here. Made me think of a recent tweet from Samuel Hulick – https://twitter.com/SamuelH… “People don’t do shit online unless it’s to their perceived benefit. Either increase the benefit or increase the perception. Or both.” I think that applies here as well. Winback communication should either communicate an increased benefit (or better communicate an existing one) — or increase the perception.