Watch What They Do, Not What They Say
Many of our portfolio companies struggle with the idea of changing something fundamental about their service, such as Twitter’s recent change from stars to hearts. It incites fury from the loyal users who believe they know what is best for the service.
I always encourage our portfolio companies to A/B test a change in a relatively small but representative sample of their users and to watch what users do and don’t spend too much emotional energy on what users say.
Twitter’s switch to hearts has resulted in more engagement with the favorite function. I would bet that Twitter had that data and understood it before deciding to make the change. You can’t make such an important change without first testing it to see what happens.
Loyal users are always going to hate a big change to a service they use every day. I recall the outrage when Facebook rolled out the news feed, which has become the central feature of its product. It was as if they had destroyed the service.
Users’ actions will tell you what they think about a change more than what they write (on your platform and elsewhere).
For what it’s worth, I loved to move to hearts the minute they did it. I feel like I favorite way more now. But Twitter would know. They have the data 🙂
Comments (Archived):
are there other microblogging platforms just like twitter out there, or is twitter the only game in town?
Yes! And it’s easier to say “I Liked” than “I Favorited”.
Friction is 4 syllables???My goodness, Twitter’s origins story was about how easy it is to update people about “What you’re doing / What’s happening?”Before we know it, someone will make Tech so frictionless we simply think something and the tech just happens.Forget clicking on ♥ signs which some users see as friction. FB wants transmit direct to our brains!
For me, it’s the animation of the heart, when you press it, more than the particular icon switch, that pleases most.I think that Anthony Volodkin (@fascinated) was ahead of his time here in having the “love” feature, ie: “Fred just loved this song”, rather than like, +1, star, etc.
animation drives me crazy. an orgasmic heart. seriously!
heh, ok it is a teeny bit weird that blood squishes out when you click it.
phew. i’m happy its just blood.
ewwww
I haven’t clicked it yet. Can you describe what an orgasmic heart looks like? I’m really curious 😉
“I feel like I FAVORITE way more now” – GOTCHA!Freudian slip??Should be “I feel like I LIKE way more now”
like is facebookfavorite is soundcloud, tumblr, twitter, etc
Not any more. That’s my point.Twitter is now ‘like’
no. its favorite. that is what i said in my post.
@liad:disqus @fredwilson:disqus — Here’s what Twitter itself suggested the heart could mean.Bottom line: it’s whatever the user wants it to be when they use it.In TWTR’s database, it’s stored as “Like” because it would be inefficient for them to log it as “could be: adores, aww, high five, hug, lol, wow, yes, read this / favoriting this / and more.”As I commented to William: “Friction is 4 syllables (fa-vor-it-ed).”
Watch what they do is a huge reason why doing fast and quick prototypes can be so helpful in product development.
Let’s get this bar fight started.1/ the star/favorite signal allowed ’emergent’ behaviour’ – something I remember USV being all over. It could serve as a read receipt/lol/personal bookmark and much more.2/ heart/like is extremely prescriptive. it forces users to use the service as centrally mandated. it further closes the platform. it assigns certain intent to a user action which is likely incorrect.3/ it retroactively and unilaterally amends the intent behind hundreds of millions of previous user actions. akin to ‘RTs’ being relabelled ‘Endorsements’. A mass bait and switch4/ i think the usage data is soft and will decline over time. even if it isn’t, a 6% increase for infantising a product seems blah.5/ not all engagement is created equal. aggregate volume may increase. but real engagement? meaningful engagement and not just meaningless platitudes?6/ the new terminology, the orgasmic heart animation. just serve to telegraph where twitter wants to head. and whilst appealing to teenage girls, a la snapchat, is probably good for business, i just find it all a bit mehEdit 1*****For the record the above is my real POV but I could just have easily argued the exact opposite*****Edit 2*****I have unilaterally decided to change the meaning of disqus upvotes They are now called disqus cooties. By clicking the button you are certifying you have the cooties.How you like them apples. *****
Twitter is not really a community and engagement is not really its metric of it’s value.Meaningful engagement on Twitter–not sure what that means.
engagement is everything. thus the change. to encourage it. no?
Twitter to me is a river of information. Twitter to someone else might mean engagement. I remember when kids were using Twitter to communicate, but it was too public and tweets didn’t disappear-so Snapchat.
river or not there’s still engagement right? that’s the name of the gameread and/or like and/or reply and/or retweet.without those, game over
I remember when kids where using snapchat because its content disappeared. Now they just follow internet-celebs. Kind of like twitter.
Touche
true.remove meaningful as a criteria.my point is that heart or whatever, they are just gestures, valuable as that certainly certainly, but not more.
TWTR might not be a community, but communities exist on Twitter. Communities exist on all platforms. That doesn’t mean that all platforms are equally as conducive to or supportive of communities, but people by their nature form groups and glomulate together, even in the noise of twitter.That said, engagement is a metric for twitter. And also for TWTR who maybe be searching for any positive metric to report, regardless of it’s actual impact.
glomulate ;)the dynamics of groups is the key metric of all all interactions. no argument. i;ve blogged and presented on this endlessly.there is a difference between gestures and connections, between gestures and conversation, between a gathering of people rubbing shoulders and engaging as a group.that’s all.
i agree about community. it isn’t a community.is ‘to use’ the same as being ‘engaged’?
Before I read your comment, I could have cared less about the switch.But you raise some excellent points.Now your position is my position. Kudos.I will now defend it until my death …or until I get my morning coffee.
.Death and the absence of morning coffee are more closely related than one might think. Do not underestimate the importance of coffee, in general, and morning coffee, in particular.JLMwww.themusingsofthebigredca…
I wonder if they tested having the Star AND the Heart.My gut tells me no.Their A/B test may not have been the A/B/C test it should have been.
Twitter is thinking to small, let every user select and customize their own like button
dont even joke! back in the day (2007) we built customised options into our ADD social games platform. terrible idea.
What sender and receiver want and perceive are totally different. Most apps/platforms don’t map out all the different pathways and contexts in which people post/send/create information and receive and distill it, and further (hopefully) repurpose it (across their’s and other’s platforms/apps to account for infinite demand/consumption). Doing so ends up with a better way to satisfy marginal consumption on both sides and not force average consumption models on users.
That’s very much like the Slack ’emoji reaction’, which is a great feature and opens tons of new use cases.
Yesss. I vote for this!
I feel your pain but definitely it’s a case of “what got you here isn’t gonna get you there.” For better or worse the heart is the standard now and if you aren’t pulling in the kids you don’t have a LT future. I got my kids their Twitter handles ages ago; they are active in Insta, Vine, Snapchat; they’re not seeing the usefulness of Twitter and wonder why it’s so awesome I grabbed their handles. A serious case of “Who cares, Mom?” {As a side note, how tweens handle handles and bios is fascinating. Super fluid.}I have long enjoyed the aspect of the Favorite to anoint both the giver and the recipient Influencer Status. It’s an intellectual ego thing and I’m totally down for that. The star plays out as tastemaker-y and non-inclusive. So it sparks conversations but doesn’t give them staying power. And that’s where we hit a wall.I agree that the retro change is awkward. I, too, have favorited things I don’t love. But I think that also underscores the confusion it causes: 8 years on Twitter and I still don’t know instinctively what to Favorite and what to RT, and what the difference really is. And so I often do both and that feels weird.The Heart doesn’t overlap with the RT. It’s — I like it! I like you! The RT is — I want to share this message. To me, it’s clearer.A final word on A/B testing. I’m sure they’ve sampled plenty and that would include lots of age segmentation and gender splits. I believe the user base is WAY more female than it was. But, women don’t get favorited, followed or RT’ed nearly as much pound for pound. (I can’t remember where I’ve read this.)So, my ovaries and I have to wonder: is this Heart thingy a tweak to better serve and more deeply engage a large and growing community which the product wasn’t initially designed for?
Maybe… just maybe… it’s because their mom got their Twitter handles. Im not sure how old your kids are, but there is a point at which the coolness of the stuff their mom gets them has significant diminishing returns.
Hahahaha!Are you saying they’re kinda like the icky faux antique rug my MIL gave us and insists is special and we must love?Guilty!
“what got you here isn’t gonna get you there.”Exactly. The accidental mondo success Twitter is now was not due to the original plan, so why/how is it expected that some of the original planners using perspectives derived from the original plan (mixed in with financial markets pressure) to map the future will produce more of the present success?
Very interesting point there.I wonder if Tw has a user advisory board (or boards). I too feel a disconnect that its growth and product development were driven by user and developer involvement, and now they want to take a top down approach. Liad’s ‘bait and switch’ sentiments often ring true when dealing with Twitter from a developer standpoint.From the outside it feels like Tw either has no idea what they’re doing, or knows exactly what they’re doing (but maybe not in a good way).
Yes, I think some of the current issues would be non-issues or resolved faster if key user contingents including developers were well represented internally.
“8 years on Twitter and I still don’t know instinctively what to Favorite and what to RT, and what the difference really is. And so I often do both and that feels weird.”That def resonates. On the other hand, if switching from stars to hearts is an attempt at appealing to women… :Maybe Moments is a great idea. Maybe hearts is going to move the needle. It’s just, I can’t help but wonder why what looks to me like some of the lower hanging fruit (lists, a way to “pin” tweets you want to remember — which a lot of people, journalists included, used Favorite for) is being neglected.It’s a fascinating product, Twitter. Means something different to almost everyone who uses it and is misunderstood by so many.
This is telling –> “The Heart doesn’t overlap with the RT. It’s — I like it! I like you! The RT is — I want to share this message. To me, it’s clearer.”If that is the case, then it sounds like they are moving closer to Facebook’s “likes” vs. “shares”. What you describe is definitely the functionality difference I see between the two over at FB.But Facebook didn’t build a big business around likes. It built it around shares. The difference in RT’s on Twitter and Shares on FB is that FB figured out a way for Shares to go deeper and longer and be more connected—-and so now shares have actual, long-term, monetary value to advertisers and content producers and those looking for a way to get messages to last a long time with a specific audience on FB’s platform and therefore have real value. If I were Twitter, I’d look for ways to make the RT go deeper, which is where I think the real monetization opportunity is.
because expanding media and causing arbitrage is a business 😉
Yes, I do know what you mean about arbitrage (and that is a huge reason to sigh). But arbitrage is getting a message out there (or more specifically, paying to get a message out there) ONLY because it earns revenue. The real winners are those who are getting a message out there because the message itself has value AND it earns money. Brand builders know this distinction. And I wonder if there is a way to look at Twitter as a system where the hearts find the value and the RT’s create additional value.
Exactly, Liad, exactly. But I’m like an old married couple with Twitter. I’ll offer my uninvited critiques, make little remarks, ignore it for a day or so, but not really going anywhere. Yet a little less affectionate because I was more generous with stars than I will be with hearts. You can’t change someone’s love language overnight.
“You can’t change someone’s love language overnight.” <– awesome quote! I’m going to steal it, and attribute you. Not sure where/how but I will. Just sayin’.
Hey YOU!
No. YOOOOUUUU!!!!!!Missed ya. Trying to get my AVC mojo back on. 🙂
Honey, you set the standard for AVC mojo. Can’t tell you how heartwarming it’s been to see you pop up on here. Of course you would choose a time when I’m just able to skim by for brief moments.
GROUP HUG.
♡ it too.
Isn’t that from the book about the 5 Love Languages by Gary Chapman? He also did a twist on that and wrote one called the 5 Languages of Appreciation in the Workplace. Highly recommended for entrepreneurs who don’t have a lot of people managing experience. My cliché is that “We build people, and people build the company.” An employee who knows that their work is appreciated is generally more productive and contributes a lot more to a startup, than one who doesn’t.
I ♡ this comment.
upvotedremember THAT change?
Upvoted.I think actually when we lost ability to see who upvoted….that was when I lost my Disqus mojo. So glad it’s back, it’s like home.
Agree with #2 – conversations on Twitter evolve in so many ways, yet all they could see was users either love it or not? E.g. Looking at Emojis shows a variety ways users like to annotate convos aside from just “liking” it or not.
to me the heart and the star are the exact same thing. i have always treated the star on twitter exactly how i use the heart on tumblr and soundcloud.i think those who take this too literally are unhappy while those who understand its just functionality, like me, don’t give a shit about this
Not to labour the point but you surprise me Fred.”Its just functionality” is the antithesis of USV. Remember software is commodity. Dentist story.Its all about the network. The community. The mind share.Defensibility in network and engagement. Not functionality.Ding ding.
its a button. it does something. it could look like a coin for all i care. i know what it does for me and i use it between ten times a day and hundreds of times a day. its functionality to me. i just don’t have any emotion about the change of an icon.
So why is there an overall increase of 6%?
it isn’t a button. Symbols have implicit meaning – and in some cases, testable implicit meaning.There is a reason why I wear smiley faces as a name tag at conferences. Symbols have implicit meaning
a) ! vs !!!, b) ? vs ???, c) ?? vs ?!, d) ; 0 vs : 0
@liad:disqus makes a valid and orthogonal point though. Yes it is just functionality but it does have a semantic contract associated with it. To @ShanaC:disqus’s point below -> twitter did a poor job of defining the emotional contract so the we (the users) did. And an emotional contract is heavily influenced by the symbol used. I used the “star” for a variety of reasons: ack | bookmark | polite end of conv, etc. By switching the contract on me without any serious engagement, twitter sent out a mixed message to its avid users and early adopters.Note: i am actually not put off by this. I continue to use twitter as before. As a product manager though this is a real world lab experiment for me: I am surprised that a lot of people seem to think/assume that this reaction is unexpected or overblown. You did not say this but I sense a “meh these people should get a life” tone in your response above :). I could be reading too much into it though.
I don’t get why they didn’t just add the heart.Star = this is well doneHeart = I love thisNot exactly the same functionality, but certainly not worth the time some people have devoted to it.
Proposed list of buttons: Star, heart, like, mildly amused, completely offended, and coin (i would pay for this?)
“it could look like a coin for all I care.” You know Liad’s going to use this against you at some point in the future Fred… the star vs. heart is an amazing episode of Silicon Valley waiting to happen.
Twitter’s move is a very interesting one wrt A/B testing the architectural adage: “Form follows function.”The form has changed (star => heart) yet the function stays the same.If a building suddenly has a new edifice (it’s painted pink instead of white, for example) HOW does it affect people’s perceptions of WHO the building is for and the building’s purpose?It would be great if Twitter would share that qualitative data rather than the quantitative, “It increased activity by 6%.”Form is the art. Function is the science.My personal view is that “Form follows function” isn’t the way forward. Rather it’s “Form and function are symbiotic. When one changes, so does the other.”
It’s not a button. It’s an editorial comment.
Every time my husband decides to make a change around the house for sake of *functionality* we tend to disagree on the outcome. Functionality is not isolated.
Twitter needs Wall Street to ♥ it alongside new and existing users.For purely self-interested reasons, I hope Twitter team steers the ship in better directions than merely repainting the hulls with ♥ instead of ★.Developers also need to believe in the direction of Twitter ship and that whatever they’re building with Twitter’s API and toolkit will last the course.
that you can argue the opposite so easily isn’t a good sign
your points are excellent and noted, though i think you may be underestimating the value of greater interaction — especially at twitter’s scale, and especially when that interaction is more narrowly defined. the narrow definition makes it easier for twitter to proactively use that interaction to improve the product and make recommendations so as to enhance the discovery experience.
+1 for the bar fight metaphor.While this is a small change overall, and Zuck has shown that he is generally knows what FB’s users want more than the users themselves, Kevin Rose and Digg showed how NOT to introduce large changes and why incremental changes are generally better.
If I understand this correctly; using stars captures the correct user intention, while using hearts amends the action into unintended behavior? I believe Twitters motives for change here were sincere – more about feel good user engagement (6/supports this view). Perhaps trying to define meaningful engagement this way is like trying to define degrees of sincerity.Ultimately, what’s good for users, tends to be good for business and the brand.I really enjoyed this mind twist!
I remember being one of these angered loyal users when Spotify deprecated their “Starred” feature, in favor of the “Check”.
the star was always ambiguous, they should have changed it ages ago, never mind whether they had the data or not, in this case they didn’t need it. I reckon someone somewhere in Twitter was just being stubborn.
I love the new Hearting on Twitter, but the splash effect is a bit overkill, no? I wonder if they A/B tested that too 😉
Here’s …. Twitter mentioned at Ethereum Devcon!
Transient Datagrams is a great band name.
How’s about Transient Vicarious Aliens? Haha.
Who gave this prezi and can you provide a link? Thanks!
Presentations and video should be here: https://devcon.ethereum.org
Aargh!!! I just saw an example of Blockchain repeating flaws from old Internet.Dapp Store launched today. What did Vitalik say in March 2015?“The idea is creating platforms that do things like Uber and Airbnb but without the Uber and Airbnb,” he told me, as we chatted next to the Crypto Castle’s kitchen counter.Buterin explained how this could work. Today, he says, we rely on Uber and Airbnb to run one-to-five-star rating systems for customers and suppliers, so that we don’t accidentally get into a car with a psycho or rent our apartment to a kleptomaniac. Buterin thinks that these functions could be replaced with a Bitcoin-based rating system that would work by making all of these systems fully transparent, and available for everyone to see.”* http://fusion.net/story/471…What does Dapp Store use? 5-stars. Did they get YouTube’s memo from 2009?!!!WHY did Twitter move over to ♥? Err … because developers know rating systems (in particular 5-stars) is broken.They can’t invent their ways out of it, though, because … it’s what they know from Old Internet and it’s convenient (if imprecise and incapable of measuring ambiguities and nuances contained in the N star):* https://www.quora.com/Is-th…* http://www.wired.com/2015/0…* https://gigaom.com/2011/07/…Facebook and Buzzfeed know this so they moved to emotion buttons and Twitter is now simply copying them.My ideas for emotion buttons was shared with RCS Mediagroup THREE years before Facebook went to emotion buttons in Oct 2015.Twitter’s ♥ is like baby being able to stand upright whilst Facebook’s emotion buttons are like baby being able to take a step forward, imo.
So much for no-Iliad day? ;)Linear rating systems do do have that dot-product reduction issue; they have their place when it’s clear what that rating is for (the people asking “should I see this movie”, for example, have already selected for taste in genres, and are then filtering on acceptable quality).It’s a depth-first rather than a breadth-first approach, so to speak.
Agreed. All of this Material Design animation is starting to get to be a bit much. Loved Material at first but like most animations, they end up getting in the way. I’m sure Twitter will discover this as well.
I think they just applied the same animation that had been applied to the star, which was a thing that lots of people loved.
Do what I say, not what I do… Oh, snap!
Same with “venture” investors: watch the risk profile they invest in.
I’m curious to know which demographics that increase is with — and if there’s an increase of 10% with some demographics and a 4% decrease by others.
but is this change going to move the dial on Twitter’s growth?
“The heart was made to be broken” if the NY Mets do win the World Series next year, Im betting on the like button is changed to an apple.
If they Cubs win it, can we change it to a Bear?
http://pointsandfigures.com… I am doing a thing for Veterans Day if anyone here wants to have some fun and support a good cause.
Just imagine how much less bloviating and fewer soundbites we’d see in other arenas of life if everyone just had the sense to look at and understand the data.#SoapboxWednesday
I understand the biotech people are working on that; they’ve found I think the method for getting to “everyone”, but I haven’t heard anything about progress relating to sense (I suspect ethical concerns in testing hypotheses there).
Yes. Agreement on what exactly sense is would be the first (and likely deadly) point of failure.
All data is relative 😉
all perspectives on data is relative, the better question is did you gather enough to make your perspective less relative?
you need to have algorithmic thinking, otherwise data is just another piece of a GIGO problem
What are the community’s thoughts on survey data in general?
The problem with this approach is twofold: many times a change of feature makes real sense only over a long period of time or scaled rolled out, which means that A/B can t be helpful enough. The other is that many times the users reaction is not necessarily a good indicator. A new feature can be a good one, except it is too new or not well executed to be appreciated.Decisions about product change have to be also based on gut feeling and intuition beyond what users think or do in front of a change
Once again, the average, not marginal usage wins out. I don’t really care to heart something when all I want to do is index it for later consumption, say on a PC where I can open up a lot more links rapidly and then “be engaged”. Give me a personal star (or file) button that doesn’t connote that I like it. Twitter is just so superficial in the moment, I don’t know what value can be truly derived. Certainly very little intent and real knowledge creation and discussion can be gleaned or generated. Just banter and shout-outs where very little meaningful work can be done and value captured.That said, it’s a useful tool to create lists of subject matter experts that I can go back through later and do some deep dives when I need to rather than a google search which is ridiculously tedious and time-consuming to find like info. To me that’s the engagement and long-tail value which Twitter is totally missing as it tries to be a platform that is relevant in the moment. If I tweet something about broadband or wearables or somesuch, it would be interesting to know if someone starred the item for later consumption and opened up the link (which I invariably include).
Meh. The recently observed 6% more hearts/likes action could be due to a temporary surge. E.g. trending topic(s) with outsized engagement (compared to average rate for such) or coincidence of higher than average signups rate within testing period (for whatever reason).While I don’t get the user outrage (the star was equally bland fer gawd’s sake), the worst thing to me about this change is the lack of imagination. This was an opportunity to create something more original and in line with the platform’s social atmosphere and visual elements. If they couldn’t come up with anything, they should have considered working with some of their most active user segments – I can’t imagine what kind of cool results that might have produced. But I guess the most exciting things about Twitter are observed around/from its users, not from within on the platform or company.
“the most exciting things about Twitter are observed around/from its users, not from within on the platform or company”Yes. For example, the Moments (while I imagine are great for some people) interest me not. They aren’t conversations between fascinating people I wouldn’t otherwise get to “eavesdrop” on, and that’s what I personally love about Twitter. I learn SO much from Twitter, and it isn’t from accounts with millions of followers.It feels like the Tw leadership are missing the conversational beauty of their product. Moments are a broadcast. Liking is a one-way, one-time interaction, etc.
most of the US’s taste is bland, most of the world is bland – we live in small worlds with doses of randomness
Yet, Twitter is arguably one platform where bland increasingly collides with non-bland, so, you’d think it would be reflected in the interface too.Plus, its North American user base is apparently decreasing, with most user growth originating elsewhere like Asia-Pacific and LatAm.
My dad used to always say the same thing about dating – it is great advice. Listen to what people do, not what they say.
Stock performance year to date:Facebook is doing and is up 40%. Twitter is saying and is down 26%.Hopefully, Twitter has grander visionary plans than turning my 42,000 Twitter Stars/Favorites into 42,000 Hearts/Likes.
Yeah, great point. So maybe replace up/down arrows under member postings with hearts or some other emoji! Could be refreshing…just a thought.
It’s interesting you embrace the heart but still call it a “favorite.” They also changed the name to “like” you know :-).
“Favorite” and “Like” mean two different things. It was very uncool to change what users opined about tweets in the past after the fact. It’s a fundamental break in trust. If they can go back and change this then they can go back and change anything.This is changing user’s editorial commentary, not merely changing UI.
When contributors have excahanged offline communication via emails or phone calls why do they become lazy and subject the entire board to the re-socializing. Just call each other.People who throw kisses are hopelessly lazy.Bob Hope
I starred that
Novelty effect
Slack launched it awhile ago and apparently it was popular. I think it was Path beforehand that had a similar implementation.The interesting thing is it might actually benefit users, however it may not benefit the metrics Facebook is focused on optimizing for.
everyone cares about that.
Haha, Jim.Sorry, it’s my “Don’t write the Illiad” day. I am watching “what they do, not what they say” wrt Twitter and 3rd party developer tools.Watching what share price is doing too and what Wall Street is saying.
reading data and thinking about people–novel thought.data is meaningless, interpretation is everything sounds like something that bears repeating.