For me, the biggest downfall of IFTTT was how many services started locking them out of useful hooks. Early on, I felt like I could do anything in IFTTT with any platform that was on there. There were so many hooks available as triggers, but as time went on and services removed many useful hooks, it lost a lot of value.
Spotify is one that comes to mind, originally there was so many triggers on Spotify that were useful, but now there is only like 2 boring triggers. It really killed the IFTTT platform when there was less options from so many of the services.
They went from having less services, but a very deep amount of triggers/hooks to hundreds of services with very little depth. From a deep pond to a wide shallow ocean. And I'm not sure if that's IFTTT to blame, or the services themselves just doing it and IFTTT had no say. Could be a bit of both, who knows. Or maybe services made exclusive partnerships with other services like Zapier, and made their triggers exclusive to Zapier for some deal they put together.
And just to add, at the end of the day, anything automation related that I was doing for professional work was always done on Zapier because they had far better/more options. So IFTTT had always just been a personal playground, noting serious, which is all more reason to not pay for it.
> Worries me that with post-smartphone technology it’s more valuable for companies to remove interoperability than to foster it.
I've been in this position before. We wanted to open up our APIs, encourage integration with 3rd-party tools, and make our product as open as possible.
Good intentions, but the unintended consequences are significant. We expected the openness to drive more sales, improve customer satisfaction, and generate more goodwill around our product.
In reality, less than 1% of our customers used the API at all. A portion of those who did were the most demanding customers we had, constantly complaining on social media that our API didn't support everything they wanted from day 1. Ironically, the most vocal API-using customers were more negative than positive for us. The extreme fringes of the DIY hacker communities can get very entitled and ugly.
Creating and maintaining the API was more work than we anticipated. With every new feature we had to make the decision to exclude it from the public API, or spend 50-100% more time integrating it into the next public API release cycle.
Overall, it doesn't make business sense to create an API if it will only be used by <1% of your customers. The only time an API makes sense is if the API will really, truly, genuinely be used by a significant portion of your customer base.
As techies, it's easy to forget that tools like IFTTT aren't mainstream outside of technical products. They're actually extremely obscure for the vast majority of the general public.
Just curious, was there any sort of poll taken of your user base prior to implementing the API to determine a rough idea of how much desire there was for it?
In my experience (B2B) nearly 50% claim API access as critical during the sales cycle, which poses huge obstacles to sales. Then of course, only 5% implement and it'll turn out their unspoken use case isn't supported. This hits home for me.
We've been there before and history is just repeating itself because people are not insisting on decentralised technology and open standards. They've forgotten (or have never known) the dangers of lock-in and are giving in to the allure of fancy web services. So we will burn ourselves again, until it becomes untenable, and then we will go into another cycle of decentralisation.
For social things, decentralization is always going to be a non-starter. I'm never going to be able to convince most of my friends to choose Mastodon over Twitter.
Things aren't sufficiently bad yet for the average person to consider doing this since the present way works well enough for them. These things happen gradually in the beginning, but after a critical mass of people have heard, considered and tried an alternative, it can become very quick.
I don't think decentralization is inherently a non-starter and I don't see a convincing argument why it would have to be. It's a matter of marketing and UX.
I actually think the hyperlink on Instagram thing is a conscious decision to limit people doing stuff like just posting news articles, and is a critical part of what makes Instagram feel very different than other platforms, in the same way people sometimes claim Twitter's decisions on short content affects its community (for better or worse). Like, I 100% agree with your point of companies limiting interoperability, but just not that example (and actually maybe also don't agree it has to do with smart phones: I bet a lot of this walled garden BS would also happen if everyone stayed on the web, and I feel was in fact already happening before apps became a thing).
It seems to me that Instagram is trying to limit automation and linking to stop it feeling businessy. While some automation platforms exists, scheduled posts break all the time, it's hard to share responsibilities for posting, and all that. You can't even upload images by logging in to Chrome unless you hack user agents or something like that. They want a one to one mapping of account to person and have you use the app when you're out and about instead of sitting in an office, for a more personalised feel. Probably because we already have Facebook for that.
Its the age old walled garden approach isnt it? Capture the market.
We aren’t the customers and we need companies that will treat us properly, but imho the root cause is providing everything for free - that will necessarily need to change.
Are you saying that "scale of spam" is a strawman? I'm stuck on this issue, because GP used the term "internet feature", and the OP claims that they can be abused: reductively this becomes, "Are links good or bad?"
If you browser instagram comments for more than 10 seconds, you have a 100 percent chance of seeing a bot or a thot. They do not give a fuck about spam. Only engagement.
I'm guessing he's talking about thotbots that go to "famous" "influencers" pictures and post comments to lure them over to said thotbot's owner and raise their magic internet numbers as well
Part of the problem was that IFTTT started charging businesses a few years ago to add support. I think they used to charge per trigger, so lots of companies pulled everything but the bare minimum. Now IFTTT is trying to charge consumers instead but it’s likely too late. A cloud-based consumer automation platform is of dubious value in a world filled with smart home platforms and local options like iOS Shortcuts.
> Part of the problem was that IFTTT started charging businesses
Evidence: I run a SaaS and had a long-standing TODO: "Implement Zapier and IFTTT support". Well, I eventually did implement both, and then learned that in the meantime IFTTT changed its approach, started charging businesses, and the amounts are not insignificant. I had no idea if any of my customers actually wanted IFTTT, so I simply killed the integration and sticked to Zapier.
Both models can coexist (Zapier charging the end user instead of integration partners, IFTTT offering for "free" or very low cost if integration partners will pay and absorb those costs elsewhere, maybe baked into recurring revenue or the cost of hardware sold), but the use cases are drastically different.
I want to say our tiny startup was quoted about $30k? Then they dropped it recently to $199, and now it's free.
I'm guessing "hope Google calls asking for a new integration and charge them a million dollars" was the business model, and they're hoping to transition into a model where they charge users.
> For me, the biggest downfall of IFTTT was how many services started locking them out of useful hooks.
Absolutely. The 2000 - 2010 time range was filled with such great hope for APIs, for expanding humanity creativity.
The last decade though has been shuttering & closing & withdrawl of computing, the systems receding, getting further & further away from general usability, a retreat into the walled garden, systems effervescing from manipulability, going up, into the cloud.
I've been calling the before time the Pax Intertwingularis era, the peace of intertwingularity, interoperation, interfacing of systems. When we were all excited to build & interconnect & share, when the hope, what we all saw, was humanity getting better & better from the power to wire ourselves together in new & changing ways.
We basically allow you to create your own custom triggers based on any change happening on a website, with value presence, absence, increase/decrease detection (deltas) etc.. and we’re compatible with both Zapier and IFTTT (and custom webhooks if you want)
When automation gets too much traction with social media services, it makes sense that they'll pull the plug as you are avoiding being monetized with your attention on their platforms.
I actually Zapier has the same problem: mediocre depth. The customizability on logic is weak compared to newer competitors like Integromat, Tray, etc. I hope Zapier takes heed.
Spotify is one that comes to mind, originally there was so many triggers on Spotify that were useful, but now there is only like 2 boring triggers. It really killed the IFTTT platform when there was less options from so many of the services.
They went from having less services, but a very deep amount of triggers/hooks to hundreds of services with very little depth. From a deep pond to a wide shallow ocean. And I'm not sure if that's IFTTT to blame, or the services themselves just doing it and IFTTT had no say. Could be a bit of both, who knows. Or maybe services made exclusive partnerships with other services like Zapier, and made their triggers exclusive to Zapier for some deal they put together.
And just to add, at the end of the day, anything automation related that I was doing for professional work was always done on Zapier because they had far better/more options. So IFTTT had always just been a personal playground, noting serious, which is all more reason to not pay for it.