docker / roadmap

Welcome to the Public Roadmap for All Things Docker! We welcome your ideas.
https://github.com/orgs/docker/projects/51
Creative Commons Zero v1.0 Universal
1.73k stars 251 forks source link

Advanced pull analytics #162

Open kencochrane opened 3 years ago

kencochrane commented 3 years ago

Tell us about your request I’m looking for something like google analytics for my Docker images that are pulled by users from hub. These would be public images that anyone can pull.

Today you can see how many people pulled total but it would be great to know.

Having this information could be a feature that you only get if you pay for and that would be fine, since not everyone would need it.

Helpful for companies who are distributing their software via hub.

Which service(s) is this request for? Docker Hub

Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard? I want to learn more about the people who are using my Docker images so that I can make a better product. I can only see number of pulls, and that isn’t enough, give me more. I have talked to other people about this in the past and it is a feature they are willing to pay for.

Are you currently working around the issue? No, you have the data, so I can’t do anything.

Additional context This could be another reason to have people upgrade to a paid plan, or if you want you can build it as an add-on for the folks that really want it, and charge more.

nebuk89 commented 3 years ago

Hey @kencochrane thanks for the input! What sort of demographic would be interesting to you 👀 (the others are pretty self explanatory 😄 ) As far as anything else 🤔 I would imagine that would roll into demographics really for the main part

kencochrane commented 3 years ago

This is the stuff you probably don't have available especially for the anonymous pulls. For authenticated pulls you could gather the data by adding more detailed info on peoples hub profile pages. If they fill it out then you get it for free.

Another way would be to have people link their social accounts to their hub accounts (maybe allow people to login with those social accounts (oauth)) and then you have access to more info on those users for these answers.

Third option is to buy the data from a 3rd party like https://pipl.com you put in an email address and it returns some of this info for the person.

Of course you would have to update the privacy policy letting people know you are doing that, and also maybe make a way for people to opt out, etc.

Since you are aggregating the data together in a larger dataset and are not singling out just one person hopefully it doesn't go to low into privacy.

You could go one level further and develop high level generic buckets and then just place people in those buckets for example for occupation/ Role below you could say (student, engineer, management, other, etc) and then put people in those buckets instead of having every single occupation / role there is.

Here are some ideas, I'm sure other people would want similar or different ones.

Other people might care about more traditional demographics age, gender and educational background (college graduate, etc) best to ask around to see what people think.

I would say start with the data you have, and then add these advanced demographics based info later, you might find out it isn't as important and if it is maybe people would pay more and then you can offer super advanced analytics that you can sell for even more money.

Hope that helps, I would be interested to hear what others think.

ximon18 commented 3 years ago

The above is useful but only for people pulling images. I imagine (most?, many?) pulls are by continuous integration systems or production deployments that don't use a local registry. For these the account pulling might (should?) be an automation account for which demographics are less relevant, but if possible it is useful to separate those out in metrics, and also if using your own images it can be useful to exclude those pulls so you can see how many real customers and customer pulls there are. It would also be good to have metrics per tag so you know if customers are upgrading or not.

amyb12345 commented 1 year ago

@kencochrane Would our Insights and Analytics tool meet your needs? https://docs.docker.com/docker-hub/publish/insights-analytics/

If not, we'd love to hear more about your use cases and what you'd want to see.