Good morning 👋
I’ve spent the last six weeks or so on the dating circuit. My strategy is to go on as many to increase the probability of finding the one. Some didn’t get past the first date. Some are on like the 5th and we’re still trying to figure out whether we’re a good match. All in all, there’s a few still in the race, and I’m actually considering going with more than just the one. My wallet won’t be happy, but it has its benefits. In case you’re wondering, my wife gets this email too and she’s actually been quite supportive.
I had no idea it will be this difficult to find a new & suitable partner… who can reliably report on the weather, of course!
I first picked Dark Sky many years ago when Appy Weather originally launched for two reasons: 1) their clean API and, most importantly, 2) their independent developer friendly pay-as-you-go pricing model. This combination made them an obvious choice to partner back then and still to this day. Pretty much all the weather providers I’ve been in conversations with expect a flat fee up front every month that gives you access to a specific number of calls. A much more expensive commitment.
What I’m trying to avoid is paying for something I don’t end up using i.e. I pay a flat $500 monthly fee to weather provider X for a few million requests, but only end up using less than half that. The sort of flexibility needed to reach this result is unlikely, and so the active goal is to minimise unused-but-already-paid-for requests as much as possible.
The situation isn’t helped by the overall uncertainty on the number of requests I expect to make to the new weather providers. See, I’m keeping Dark Sky as a provider till their API shuts at the end of 2021. I’ve had so many ex-Dark Sky users move over recently, it would be idiotic for me to do otherwise. They’re also staying as a free option, and most likely as the default because they remain the most sensible choice.
What this means is whoever I pick will definitely not get the majority of requests per day. As the options available are mostly (much) more expensive, in order to get my money’s worth, I could make my choice the new default – this would ensure the majority of my expense is to a single provider, rather than two because most of the requests were with Dark Sky and I had to pay a flat fee to a second regardless. Or, and this is what I’m leaning towards, I offer the new provider(s) only to paid users – because there are less of them compared to free users, I don’t need to pay up-front for as many calls. Following still? 😂
If I do go with the second approach, based on the numbers I’ve been quoted, it’s still unlikely the current $3.99 annual subscription fee would adequately cover the added expense. The thing is, regardless of these developments, the plan this year has always been to offer a more premium IAP (Appy Weather++). And so it may be that any of the new providers end up being available to these users only. There won’t be as many, which is actually good, as that means less requests, and with a more expensive price, there doesn’t need to be as many of them to at least cover the flat monthly fee, to start with anyway.
The longer term goal is to have as many providers as possible to ensure everyone has at least one suitable choice regardless of their location. But before this can be possible, I need to make sure the limited budget available is spent wisely with my first pick(s). This is why this has taken longer than I like, but I’m hopeful to integrate the first provider later in June. Although there’s no immediate rush given Dark Sky’s API is available for another 18 months, I feel it would give confidence to all users if there are alternatives added sooner rather than later to emphasise through action that the app’s future isn’t tied to Dark Sky’s (I’ve had to clarify this often since the Dark Sky news, leading to the FAQ update).
In anticipation of future agreements, base integration work is mostly done, and what remains is getting everyone data’s to speak to my custom logic. There’s a lot of it, so this is a big job. What doesn’t help is not everyone offers the same data set, so the logic has to be customised to work with what data is available. I’ll most likely be spending most of next week on this at the expense of iOS dev (this is what I focused on this week, and although I feel it’s months away still, I’m starting to feel really good about its prospects).
This has ended much longer than I expected, but thank you for being interested, because writing this all out has helped gather my thoughts on the matter.
Have a great worry-free Sunday!
Best,
Bardi
You have a wife and are dating other girls? Or did I miss something 😉
Which providers are currently on your list?
Is it possible to have two subscription levels? I'd be willing to pay $1 a month or something.