https://apify.com/ logo
Join Discord
Powered by
So what would happen to the quality
# 💻creators-and-apify
a
So what would happen to the quality score where tasks are involved??
e
I hadn't considered that - I'll make sure it's not a factor anymore, otherwise idk how anyone with 100+ Actors can optimize for it.
a
I was just wondering... It doesn't make sense to recommend creating tasks (which is a good practice itself as it enforces a dataset schema (for predictable actors output) that many devs don't like to have - regardless of whether it was generated using AI or a script), but then cap task creation with a relatively low limit. That means you can only optimize up to 100 Actors if each has a single task, or 50 actors if each has 2 tasks etc... I think both are low limit if you're targeting different various long-tail keywords queries... an open question..
i
What is the current limit? Is it a total per account? Or a total daily?
a
i
100 per account makes no sense
We were instructed to improve our actors by creating these tasks. I have developed workflows myself to create those, and now I just can't use them?
Whatever the measure it is, to be fair, it must be per actor, not per account
x
Apify has tons of fancy params leading to actor rating but absolutely basic which is speed of actor is missing. This is basic feature available eg on rapidapi. Actor rating is useless if the actor itself is slow as hell etc. As for tasks I noticed they are totally ignored by google thus unusable for seo at all. Will be happy if you prove me wrong.
a
I'd say yes, but speed is subjective. Plus, the compounding effect comes directly from popularity and user stickiness - users won't stick around with a slow actor, which directly impacts MAU and, ultimately, the popularity score that feeds into actor quality. (This speed suggestion came up during one cantina call, but the argument was that not all actors are scrapers & the team decided to set it aside..)
i
Do I need to choose which actors can have a task and which not? This makes no sense.
If tasks were a bad product decision, maybe it is just better to deprecate it for everyone.
The Actor quality page is currently suggesting that I add a task to 45% of my actors... but I can't 🙈
l
> If tasks were a bad product decision, maybe it is just better to deprecate it for everyone. They are good product but it generated too many pages influencing SEO and attracting spam. I hope we can refine this over time, I agree 100 per dev is quite low for large publishers.
Regarding Actor quality, this will be replaced a bit with AI over time. We already deprecated some metrics that didn't make much sense like "success rate". Data response time also doesn't make much sense since the response differs in each Actor. Only AI can reasonably analyze if the Actor works well overall. Ofc it will work together with current metrics like active users, revenue, issue response time, reviews etc.
i
Okay, so you have plans to improve this in the future. Perfect. But what about today? You have capped the number of tasks at a fixed maximum of 100, with no prior warning or anything. Tasks affect Actor quality, but I can't create new ones, and I assume many other developers are in the same situation. I'm now at a disadvantage compared with other developers simply because I've worked to build a broad portfolio. I can't create these SEO pages for any new Actor unless I hurt existing ones by removing their tasks. I know I usually have strong opinions. But honestly, this looks like a really bad decision. It doesn't feel fair, and it penalizes legitimate developers who have been working hard on the platform for years.
If your problem is spam, then work hard to remove it. Don't hurt those who are trying to do things the right way.
p
I am in the same boat as @igolaizola here. I followed your advice from a few months back on creating tasks. Invested a lot of time in getting the workflows smooth, published over 500 tasks, and now I can not publish anything more.
x
I would disagree. If someone looks for google maps scraper which you have thousands of in your store, ONLY relevant info would be scraping speed. Apart from apify pushing their own scrapers in front of the store (which I absolutely understand, it's your website, your products, your rules so let's call these scrapers Featured) there is no way for user to choose fastest one, he will probably stick with any because he never finds out he could get 100K rows in 1 minute instead of 10. The same with other major scrapers. Absolutely don't get why there couldn't be Sample Run Speed info. Pretty easily doable using AI agents imho
l
I already reported the actor quality issue yesterday. Thanks for the report
We aren't pushing Apify-maintained Actors unless it is some niche Actor we want to temporarily market. They have simply been there from the start when we had to bootstrap the Store from scratch. Apify maintains only few most common sites so we have hands on experience in maintaining Actors. It doesn't make sense to measure speed generically since there is no equivalent between 2 outputs. Once we have a good AI evaluation, we can add it as one of the metrics.
i
What about this part? > I'm now at a disadvantage compared with other developers simply because I've worked to build a broad portfolio. I can't create these SEO pages for any new Actor unless I hurt existing ones by removing their tasks. > > I know I usually have strong opinions. But honestly, this looks like a really bad decision. It doesn't feel fair, and it penalizes legitimate developers who have been working hard on the platform for years.
l
reported too
PreviousNext
Powered by