Star to heart. Sunny yellow to shouty scarlet. Favorites to likes.
I’ve got 4,652 of them, and there’s plenty of things in there I don’t *like*. Does it matter? Yes, because Twitter’s replaced the meaning that I had given to my use of a label, with a new one.
Users have lost the benefits the neutral favorite star gave humans who can read and interpret nuance, sarcasm and humour, that data analytics tools cannot. The change has deleted context and flattened every emotion to a ‘like’ regardless of its presentation. Twitter has reduced the range of meaning given to a favorite with a sentiment that is on or off in a ‘like’.
To me favorite was a verb, a non-value based indicator of preference that I could action to indicate a range of interaction in context, with or without sentiment. Twitter seems to have seen it as a noun which allowed only one, if it believed “you might like a lot of things, but not everything can be your favorite.”
This sums up the change to me. For users Twitter is about action and interaction. To Twitter it’s about stuff.
I suspect a change that is more geared to machine reading and being able to use the data with more confidence by commercial third parties because ‘like’ is more defined than the range of meaning users assigned through a star and favorites in the past. It will be interesting to see what change it brings in big data analysis in months to come.
So while what it looks like matters for its use (especially with professional or sensitive content) it’s how the click is interpreted that adds the like, that is more important than how it looks on the screen. And we’ve lost some value in how we can use it.
I’ve three suggestions. Two for Twitter HQ and one for developers.
Dear Twitter, if we have to have hearts and likes, can we at least have them default to our own colour scheme and not impose SHOUTY SCARLET please? And we need a bookmark tool. Thanks.
and Developers? If you can create an app that would give me control to mass un-like my tweets to date, I’d use it. I’d even pay for it. And like it.
Not because of how hearts look to others onscreen. Not because the word like has overwritten my previously starred favorites, although both these things matter for perception of use going forward.
Why I want to change my past favorites to undo their relabelling as likes is because although what the icon looks like is irrelevant, (you can have a heart, a star or a flaming taco if you prefer through tools like Stylish or Starback) it’s how Twitter is purposing the data it collects through the use of that function and how they are used by others that matters.
By adding the click that is not RT that-shall-not-be-named (favorite or like as you choose to call it) you add data to your profile of use that Twitter uses and is used by third parties.
And Twitter just relabeled that.
And all the things you or I may have previously bookmarked with a star, may not be things we like, yet someone else may interpret that data to mean we do. This could be more important if web browsing is to become accessible by the state and would concern me how generic ‘likes’ could be misinterpreted. Especially in other countries.
It feels rather unfair to attribute a meaning that I did not give to historical content. The shift in meaning that has been attributed to what I did in the past is outwith my control.
While I can control what I add a like to now and in future, having lost its range of empathy, it offers less value to me as a user than favorite but I suspect it is probably more valuable to Twitter. I feel this change is of more benefit to them as an organisation for simplifying behind-the-scenes analytics using my data.
Twitter as a tool offers value others cannot, so a new function to acknowledge and bookmark, would be valuable to users and nice to have to make up for feeling more ‘monetised’ and more value to Twitter. Our data and advertising base are valuable in this ‘free product’ transaction between users and the organisation.
Will the changes otherwise affect users behaviour? Probably.
While I will always ‘like’ dancing baby stingrays. I will choose now what I ‘like’ more selectively and use it less widely in future.
And I’ll consider making correction by removing the ‘likes’ attributed to my favorites. I really dislike inaccurate data where it can matter much more, than how it looks on Twitter.
Granular definitions matter and was useful in the neutral ‘favorite’, because not all ‘likes’ are equal as I’ll go into more detail in examples.
Here is what I’d like when I use Twitter: transparency, choice and control.
More detail on why it matters and examples illustrated in ten tweets below.
 web browsing could be misinterpreted http://www.theguardian.com/uk-news/2015/oct/30/police-seek-powers-to-access-browsing-history-of-uk-computer-users
 https://nandosigona.wordpress.com/2015/10/13/the-politics-of-counting-people-at-borders/ on the importance of accurate counting of people in immigration
Why does the difference matter?
What does that information in my user profile tell others and who is using it for what purpose?
There are reasons why you or I may have wanted to ‘favorite’ and not retweet. Favorites expressed some sort of value judgement, that a RT does not. You may agree or disagree with something you retweet and there is no way to tell that apart. But by adding a star you added some sort of layered meaning, associated with your twitter handle. And meaning is dependent on perception.
Since the change has been backdated, anything I previously favorited is now liked, and open to misinterpretation. It may not matter that much to you or me personally as an individual, but consider journalists who have ‘liked’ a series of tweets about an event for reference, from a country of an opposing nation or group in war time. How might that be interpreted by authorities?
I dislike it for two reasons and it leaves me asking questions:
- as Twitter reduces nuance and aims to cater to big data for machines: Who is Twitter for? Users or users of our data?
- I’ve become aware over the course of the last year of the use of the data I don’t see in Twitter and that those data can be used by other people. I didn’t know when I signed up and until more recently, that analytics could and would study my tweets’ content, location (if I used it), time of posting, reach to others, monitor and analyse my use and interactions. Tools like this. What is Twitter used for?
- Those invisible third parties are analysing our twitter data for unknown purposes and interpreting it. Presumably in order to act on it, to better target ads or other forms of activity. While I know that twitter is public (unless you choose to restrict it) I didn’t know how other people with machine aided tools could read more into my twitter at a glance of their dashboard, than I could from my everyday use. Why are they doing analysis with that?
So what does my data tell them?
Well, here’s the thing. It probably doesn’t tell them, what I think it would tell them. Or perhaps what they think it tells them. Because favorite was neutral its use was nuanced. And personal. Like is pretty much, well, like. Always a positive sentiment.
To answer questions 1-3 it would be useful to see a report on demand of what my data looks like to others, and examples of how they are used. Too few users understand this.
Why it matters that hearts do not equal stars
A star was most often a bookmark and more. A ‘I’ll come back later when I’ve time’ flag. Or ‘thanks, I’m leaving this convo now, see ya’ kind of sign off. Or a quiet acknowledgement.
You can’t do ‘quiet acknowledgement’ with a crimson heart that appears in a cloud of glitter when clicked. (albeit pretty rainbow glitter.) It’s more of a shout and a hug, then a quiet wave from a distance. I wonder if the test audience was all American?
Overall, a star didn’t equal a heart. A favorite is not a like. So what my user profile tells others now, has been changed by Twitter, not by me. I’d like to get this control back, through a tool to undo my historical likes.
It’s about how it feels
The sentiment in a like is flat. On, or off.
In contrast there is a kind of curve-of-nuance in a favorite, that a like cannot express. It goes from something positive through a neutral acknowledgement of fact, right up to I-strongly-dislike this but acknowledge its content or its poster’s position.
Here’s some examples, starting with the positive.
- Like means like. OK, here is a tweet I favorited. And I also ‘like’ the image of weather data. And while I ‘love’ many meteorology related things, I do not red-heart ‘love’ them. I like it a lot. I like that he posted it. I have no feelings about the poster. (I do not know the poster.)
3. I can like the purpose and the content, but heart can still feel inappropriate. This is a positive acknowledgement mixed with ‘I agree’! However it now feels very odd to have red ‘heart’ed a professional body and signalled that to those @mentions included.
4. Here’s an acknowledgement to signal appreciation – not emotion – ‘I know and agree, and add my own support too.’ 5. An acknowledgement with subtlety and an ‘I know and I agree, and a bit of irony’ thrown in for good measure.
6. But here’s where it gets tricky. I certainly don’t *like* this on pay cuts for NHS staff although I did like that the tweeter posted it. It’s more of an acknowledgement of a factual statement. It’s neutral on sentiment. This is not a like.
7. Further along the favorite curve, you get into the I really don’t like the content, but want to wave in solidarity, and ‘you’re right to stand up for this’, kind of acknowledgement, perhaps because I might not be brave enough to say so myself. The heart icon, does not work. This is not a like.
8. And then you get into the I really don’t like the content, although I do like the emotion it expressed and that you posted it. Here is where the icon, and a red one at that, does not work. This is not a like. And there are even worse examples in war and suffering.
There are practical reasons too, why you wanted to star/favorite and not retweet or indeed ‘like’:
9. favorite often enabled you to appreciate, thank, encourage to continue, be involved but decide, “I won’t retweet as it’s not of broad interest to most of my followers”. Heart icon, does not work.
10. The favorite was a way to acknowledge content without expressing a value judgement. So what about those tweets that are about “clock-building” that I’m simply interested in reading and favorited only for my own use, and no one else’s benefit. But now instead this signals to the world that I ‘like’ “clock building”? (When I don’t – I’m just curious). If all my web browsing is accessible by the state this would concern me how this could be misinterpreted.
It’s about how it reads
The interpretation of the written language is also different between favorite and like.
Like is used to convey a person’s attitude or feelings. Favorite is more of a preference, and more nuanced and subjective. I may even dare to say American use of like and favorite are different from British English users. You can prefer one tweet over another to come back and read later without any sentiment.
But a like is always positive.
It’s about how it looks
Another catch with the switch is in the icon itself – that the association with a heart is different from a star. The association with a red heart, in many places in the world, is of romantic love. My tweets that I favourited with a more neutral yellow star don’t mean I ‘like’ you or your content in that way, at all. (Sorry.)
The interpretation of the image is different. What my use of favorite was intended to mean to others was individual to those whose content I starred and in the context and timing of that comment.
My meaning has been overwritten to all read the same universal ‘like’ by Twitter, not by me. (Even if they say, it has different meanings, they are all positive ones.)
It’s now out of context and outwith the time in which the original ‘favorite’ link was made.
To get the control back of using the nuance of favorite with professional or sensitive content, a bookmarking tool would be useful, rather than a scarlet ‘love this’ heart.
All in all, it feels that my twitter data has been repackaged. Whether it’s aim is to appeal to new audiences, or make marketing easier, you can debate. Removing nuance removes ambiguity from responses. That makes it easier to segment.
Twitter’s drive to monetise its user data is no secret. But how my use of twitter may be marketed to others is not transparent.
Will I go back and un-like over 4K favorites-now-likes? Probably not. Had I a mass tool that allowed me to do so? Yes, I would use it. If someone were to design an app to allow me to do so, to regain control of what a system considers that I ‘like’, I would likely even pay for it. I’d like to have that choice.
It feels rather unfair to attribute a meaning that I did not give to historical content. The shift in meaning, creating likes, attributed to what I did in the past is outwith my control.
The ‘like’ judgement value may well be misplaced and open to misinterpretation, as well as more easily monetised based on inaccuracy associated with my profile.
It is the re-labelling for the tweets of the past whose subtleties have been wiped out and replaced with a uniform ‘like’, that I dislike most in the change.