I bet that for a lot of subscription based businesses, this is the case. I myself have a few subscriptions as a consumer and a lot of them have taken money from me when I did nothing. Just because I am either too lazy to do something about it or I just don't care enough. Either way, the business is happy to keep charging until I force myself to cancel if needed.
Now, it is also easy for a business to setup a simple recurring model. From a technical standpoint, can you imagine if the subscription billing functionality now needed logic to not charge due to inactivity ? How the hell do you calculate "inactivity" ? Just by checking login ? What if a user starts claiming that they never did anything even though they logged in once by mistake ? Do you still not charge them ? See how easy it gets crazy with various permutations and combinations. It will most likely end up in being a huge hassle for a business and not worth their time.
The epitome of this for me is RedGate's old Ants Profiler. It was the most awesome thing ever when it came out, and a day with it could get you a 1000X performance improvement in your code. But after a few days, you would have optimized everything to within an inch of its life. And the thing came with a 30 day trial. By the end of that month, there was no need to buy it any more, so it got uninstalled and forgotten.
Then one year later, you had a new project for a new company that had a performance bottleneck in a critical piece of code. So you'd download a new trial with your new company email address.
I actually see this usage pattern with S3stat (one of my products). Users will get excited for a few months, then drift away and cancel, only to pop back on the radar 16 months later. I used to welcome them back and kick off a job to run reports for all those intervening months. Now, I welcome them back and drop a notification on their dashboard with a link to purchase service for those 16 months they were away. It's a service with a real daily workload, regardless of whether you happen to check your reports that day (they still need processing so you can see them when you do eventually come back), so nobody has ever complained about this.
Contrast to Twiddla (another of my products), which only uses our resources while people are actually using it. For that, I'm happy to see Teachers cancel their account for the summer. I've even refunded customers who "forgot to cancel" for entire years at a time.
So, 500 words later, I guess the answer is that it depends on the business.
Do you ever pay extra when you get a seemingly disproportionate value for that subscription?
If you got the a la carte pricing you are looking for it would likely be more valuable to the service provider as a sales mechanism. So, it would be priced just right to get you to convert to the subscription model.
I'm not debating the merits of any of this, but trying to answer your question with given market realities.
The previous comments about AWS are the price/value correlation in practice. AWS prices will go up, but right now, it is more valuable, in a strategic sense to grow market share. I switched from Rackspace to AWS because I received credits from AWS/Y Combinator Office Hours. It was a big hassle to switch, but I also received disproportionate value, and not just because they were credits, but because the AWS ecosystem is great.
I am getting ready to launch my enterprise SaaS and pricing/value was a significant consideration. Not just to calculate how much I can earn, but to stimulate growth and prove useful to my customers - which is the key to eventually getting to the first two. At the center it has a marketplace and the pricing is based on the gross value of the transaction. All fees are paid by the suppliers. The service is otherwise free to use. So, I only collect fees when I help facilitate client business. The bottom-line is that this is how I can be of the greatest value to my clients.
I think AWS, and my pricing follows a similar model, is the evolution of the basic subscription model, which in turn was an evolution of the standard enterprise license model, which collected fees upfront with each new release installed.
See http://www.bloomberg.com/features/2016-adore-me/
I was looking at email deliverability services and related offerings (managed email marketing platforms) and I found it really hard to look at all of the offerings that made me choose a $10 a month plan or a $30 a month plan or a $70 a month plan when I really (1) did not know how many emails I would be sending in the long term, and (2) did not want to deal with the cognitive load of having yet another subscription to manage.
You know in their heart of hearts they are just hoping I sign up for the $70 plan and forget about it. I built my own email marketing platform based on AWS DynamoDB, Lambda, and Amazon SES and I like the fact that it has a cost based on what I send.
I don't want to guess how many API calls I want to use a month forever into the future or have to thing about Yet Another Subscription.
The Adore.me scam (and a few others) that you refer to are about signing people up for recurring billing without their knowledge. A customer buys an item from an online shop, which is expected to be a one-off transaction. But there is a well-hidden checkbox in the checkout flow to sign up for a "discount club" of some sort with a monthly fee. These are hidden on purpose, and hopefully will get banned.
But there is nothing wrong or dark about having recurring billing for a subscription service in general, as long as it is obvious that you are signing up for a monthly plan for something useful.
I guess the small difference with Slack is that they know you'll nearly always have more than 1 user.