The alternative is misery that is too gratingly boring to bother describing here. Managing permissions with Metabase groups and SSO with SAML or JWT makes life tolerable. You can manage authentication with Single Sign-On But once sandboxes are up and running, the amount of time you’ll save is astronomical compared to having to manage a bunch of different dashboards and filters. ![]() There is a fair amount of setup here, but it’s not that complicated, and we have a ton of documentation on how to do it. The idea with data sandboxing is you just make one dashboard, or a small set of dashboards, that all your customers can use, and they only ever see data relevant to them. You can sandbox data, so customers can only see the rows and columns you want them to see Metabase ships with tools that make it easy for people to learn about their data: people can look up tables and fields from the data reference, create shared business logic using models, and annotate trends using events and timelines. While we can’t speak about how other products pan out, our experience with Metabase is that customers just “get it”. Some people worry (needlessly) that giving the full toolkit could “overwhelm” their customers, or that their customers won’t know what to do with the data. You’re no longer the bottleneck, and you’ll never get inundated by customers’ trivial requests for data. There are many reasons to opt for full-app embedding when offering analytics to customers, but we’ll focus on the five most material: Customers can self-serve, so you don’t have to deal with ad hoc reportsĪrguably the most important: customers can drill through dashboards, or even ask their own questions of the data using the graphical query builder. ![]() And then switching to full-app embedding. ![]() It’s effectively as if you were to acknowledge that, yes, there exists an elegant, battle-tested solution to the problem you need to solve, and then instead opt to do something that takes way more effort (read “money”) and yields a far worse solution that doesn’t really work. In fact, we could arguably classify going this route as a “mistake”, or even a “bad call”. These tedious, employee-churning tasks will tie up resources that would be better spent on improving your software or service. If you pass on full-app embedding and try to hack together an alternative, you’ll not just incur technical debt, you’ll also end up taking on the nontrivial overhead of 1) maintaining the signed embeds for individual customers, 2) dealing with authentication and permissions, and 3) fielding ad hoc customer requests for data. But the return on investment you’ll get from a full-app setup dwarfs the ROI of maintaining alternative solutions (like a Metabase fork, or a bunch of signed embeds), especially as you try to keep up with a never-ending queue of ad hoc requests for data. There is an upfront cost for full-app embedding, in terms of licensing and initial operational overhead. The reason that some people hesitate when considering full-app embedding (hence this article) is because they worry that full-app embedding is more expensive from both a labor and licensing perspective, and they think they might be able to get away with juggling signed embeds with locked parameters to restrict data. And self-service meaning that all customers can create their own questions to run ad hoc queries, so you don’t have to write those queries for them. Multi-tenant meaning Customer A cannot see Customer B’s data, and vice versa. With this setup, you can offer multi-tenant, self-service analytics. Plus, with a paid plan you’ll remove the “Powered by Metabase” footer on the charts.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |