Recent searches
Search options
@mekkaokereke@hachyderm.ioMe: Hey app devs! During economic slowdowns it's really important to make sure that your app doesn't have a leaky funnel, and works for as many of your users as possible. Here's how you can use ML to make 5% more money from your existing app!
Crowd: Hells yeah! ML! *Pays attention*
Me: Here's how you can fix accessibility issues to make 5% more from your existing app!
Crowd: *Tunes out* *yawns*
Me: You can also make 5% more by making your app work better for Black and brown ---
Crowd: Booo!
@mekkaokereke if I recall correctly, I believe 16% of the US has some need for accessibility solutions so I’d push that 5% higher. And I’d wager that there is more than 5% being left on the table by not making solutions work better for the Black and Brown populations in the US alone which quite clearly are an even larger section of the population. (Overlap not withstanding) Do you have any good guides/resources you recommend on how to improve?
@tpak Yes, the number can be much bigger than 5%. I've seen accessibility improvements drive uplift of as high as [REDACTED].
Annie Jean-Baptiste literally wrote the book on "Building for Everyone." https://bookshop.org/p/books/building-for-everyone-expand-your-market-with-design-practices-from-google-s-product-inclusion-team-annie-jean-baptiste/7391065
(Bonus points to me for using "literally" in a way that won't make a pedant's eye twitch!)
@mekkaokereke, thanks for the pointer to the book. I'll dig into reading it next. And congrats on the literary achievement ;-)
@mekkaokereke they boo because the truth fucking hurts! If only people really cared more about making money than being shitty people.
I think the point is that they could make more money by _not_ being shitty people...
@mekkaokereke
@Homebrewandhacking yes that’s what I implied. They care more about hate than making more money.
@sickmatter fair. Serves me right for reading so late at night.
@mekkaokereke Nice slight of hand there. 
@mekkaokereke as a junior web dev specializing in accessibility engineering this makes me cry angry tears.
@mekkaokereke I wonder what happens if the recommendations is to use ML to improve accessibility or friendliness for BIPOC or other marginalized groups. Would their brains explode?
@mekkaokereke you don’t understand! Those don’t add buzz to the ol’ skills resume which will be worth more than 5% at the next role
@mekkaokereke crappy overlay snake oil salesmen: We use ML to pretend to fix your accessibility issues!
Crowd: TAKE MY MONEY!
@mekkaokereke back when I had just taken over as GM of Strategy Development at Microsoft, with a charter of coming up with "game-changing strategies", I met with folks from the Black at Microsoft (BAM) ERG. They described how large chunks of the Black community had migrated from AOL to Yahoo! Messenger a few years ago, the specific tactics Yahoo! used, and why MSN Messenger was well-placed to do the same to Yahoo! I was enthusiastic and ready to charge off and try to make it happen (1/3)
@mekkaokereke But the BAM folks had another great suggestion: before trying to run it up the corporate flagpole, I should figure out what happened 18 months ago when they had a similar meeting with my predecessor ... who was equally enthused, but then nothing happened.
Turns out it had been the usual corporate indecision and asking for more evidence and postponed meetings and lack of resources and other footdragging that happens when there isn't enthusiasm about a project. (2/3)
@mekkaokereke Armed with this knowledge, I tried to route around the previous roadblocks ...
...
...
and ran into indecision and asking for more evidence and postponed meetings and lack of resources and other footdragging.
Hey wait a second, I'm noticing a pattern here! (3/3)
@jdp23 @mekkaokereke
I have to ask: Was the foot dragging typical of all new initiatives?
@patrickgillam No. New initiatives that had support from key executives at the time did not encounter similar resistance. So I'd say it was typical of new initiatives that key executives, for one reason or another, didn't see as important.
@jdp23 @mekkaokereke
“For one reason or another.”
I will never have the information I would need to really know, but from this distance it looks bad that the initiative for Black customers gets short shrift.
@patrickgillam It certainly was a missed opportunity. On the one hand, there are always more potential opportunities than resources so some get missed. On the other hand ... well, it certainly fits the pattern @mekkaokereke mentioned in the original tweet.
Actually @mekkaokereke Yahoo!'s success is a great case study of your point of making apps work better for Black and brown people.
What was their innovation? Customizable avatars with choosable skin color.
Combined with a well-focused brand advertising campaign, (and IIRC outreach to some or all of the Divine Nine networks), they got amazing results in just a few months. No ML required! (4/3, oh well)
@jdp23
Re: skin color on avis...
So I love MS. Always have.
So when I got on O365 and started using it, I embraced the level of customization it seemed to have. In one of its apps it offered different skins for backgrounds - but none of the graphics came in diverse colors. Just white. White hands, white figures, white faces.
On a global product. It's the 21C, for pete's sake.
@DeliaChristina Ouch. But, not unexpected. Even though it's a global business, Microsoft often doesn't seem like it fully understands diversity.
@mekkaokereke This tracks.
I pitched localizing a platform for Spanish language and potentially opening up the associated marketplace with it to tens of millions of users and was met with blank stares and passive acknowledgement.
Localization services are relatively inexpensive compared to the increase of addressable markets.
Actually, I lied, me and the group received some prize (gift cards?, can't remember). 
@mekkaokereke if we just machine learn hard enough, we will solve these problems too. All problems are linear algebra!
/s
@sophieschmieg @mekkaokereke
I have a theorem, a theorem that is mine (have had it for years now) then when we finally get around to tweaking ML models to reduce their power consumption (deploy it at even larger scale, that'll be a big cost, right?), we'll replicate most of the bad decision-making tricks that humans use, for efficiency.
I think there's a fair number of people who get irritated at "stuff" because they don't want to deal with various sorts of ambiguity, because thinking is work.
@mekkaokereke in case I've never said it: thank you for speaking publicly on topics like this.
It prompted me to take action and ask a similar question at work.