A fascinating article by the Financial Times on how the rise of digital, in its quest to bring new efficiencies through targeting, has ignored or perhaps even eliminated what makes advertising work in the first place.
I was reading a great article over at The Guardian, Why too much choice is stressing us out, and it occurred to me that what a Product is, in one sense, is a set of choices that someone else is making for you.
(Forgive me if all of this is obvious to anyone who does branding and product design – I’m coming at this as a developer)
Think about brands, political parties, sporting teams, major institutions, hedge funds – they’re all proxies for a set of choices some large group of other people, usually experts, are making on our behalf, as well as the choices they choose to allow us. We couldn’t go through life if we had to vote on every bill, design our own cars, or choose the delivery route for packages coming to our house. In order to achieve comfort we need to give up control, and giving up control is much more likely when you trust the brand.
Comfort – the choices we take away
Control – the choices that remain
Brand – the name that stands for our philosophy in the minds of customers – how we choose between comfort and control.
Oh, Brand. Almost nobody buys the latest iPhone after extensively road-testing it. They buy it because Apple makes it, and Apple has a certain reputation – one honed through many years of carefully trading off comfort and control.
I’ll tell you right now, as a developer, I could not have built Instagram, a company that grew to a billion dollar evaluation in 2 years. Not because I couldn’t scale it or make the right technical decisions, but simply because I wouldn’t have let myself believe that people would be happy with just posting photos and commenting on them. I would have come up with a bunch of other things you could build on top of that and never realised that by adding them I was taking something away.
One of the things that has blown me away at Automattic is the quality of the design team and the respect they get. This is a big deal for a company founded by a developer, and a lesson that many other engineering-led companies never learn (or learn too late), and it’s one of the things that allows us to scale as a fully remote company without the user experience becoming fragmented.
When I work closely with a great designer it gives me confidence that we can find that balance between comfort and control, that we will live up to the promise our brand is making to users, and ultimately we will be more successful in the market.
This is a belated post about something that happened a few weeks ago, but it’s worth recording in the annals of personal history.
I’ve been wanting to participate in Cultivated Wit‘s “Comedy Hack Day” foreeeeeever, and a few weeks ago I got my chance. I rocked up with a couple of my own ideas but was quickly swept up by someone else’s pitch: a tool which would let users scan the online comments of potential dates/employers/friends for bigotry before choosing to engage. Less comedy, more real-thing-that-should-actually-exist, but still… it’s all the presentation, right?
Speaking of which, here’s our presentation:
We called the tool “Free Reddit Check“.
How it worked
Type a Reddit username into the search box. If the comments of that user are already in the database, you will be redirected straight to their report. If not, it will prompt you to connect your Reddit account to the site so we can access the Reddit API and scan the history of the target user.
Our system was pre-programmed with a list of “bad” subreddits, areas where the community typically hates on a certain kind of person. For example, WhitesWinFights is a white supremicist subreddit. Some of those choices were controversial, but 99.9% of them were fairly cut-and-dried.
We also made lists of hate speech terms. Again, few surprises here – if someone’s saying “feminazi” a lot on the Internet and you’re a woman, you probably wouldn’t want to be around them.
Once we had all the comments from a user, we generated a report containing:
- “This person is X% terrible” – based on number of hate-subreddits and hate-terms they’ve used, roughly divided by how much total speech.
- “This person probably hates X” – their main hate-target, based on the number and type of hate-speech terms used, e.g. This person probably hates minorities
- A list of their subreddits, with the “hate subs” in red
- A list of hate speech terms, categorized by the target of their ire – women, LGBTQ, minorities
So that’s it! Fairly straightforward, and with lots of caching and whatnot we were able to withstand all the traffic that we were getting from various corners of the Internet, mostly articles on Engadget, Vice, fastcodesign, etc.
We’re sad to see this iteration of the site go off into the sunset, but it’s given us lots of ideas for future projects and we were amazed and humbled that our simple idea won Comedy Hack Day 9 and brought a smile to so many faces while it lasted. w00t!
Just got some great advice from our pediatrician about bedtime: make the child feel they have control by offering 2 or 3 choices, but make sure all those choices are acceptable, I.e. Involve going to sleep.
E.g. “Would you like to go to bed before having a cup of milk, or after?”
Suppress dissent by giving them the illusion of control while having none.
Then, several hours later, I was like: “Fuck. That’s how democracy works.”
WordPress MultiSite (formerly known as “WordPress-MU”) is a built-in feature of WordPress that allows you to securely host many WordPress sites on the same code and database. It was developed for WordPress.com, one of the highest traffic web sites in the world.
By using these two tools together, we can compose powerful new site networks catering to different verticals and communities around the world.