Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Ask HN: Would you want to run SQL queries on application data inside SaaS apps?
4 points by NavyDish on Jan 11, 2021 | hide | past | favorite | 8 comments
I am looking to validate an idea with the thesis that every human knowledge worker in the future is an analyst. Specifically looking at functions like Sales, Marketing, Talent etc. and their interactions with tools that do the trick for their functions eg. Hubspot, Intercom, Salesforce, Lever.

Core Idea: Trying to imagine if these functions would want to use an IDE/SQL workbench like option available inside these apps. Essentially exposing the underlying data stored for my organization. Theory is two fold:

1. Admin dashboards are never adequate 2. Customized dashboards are cumbersome to set up (Eg. Hubspot) 3. Data crunching from SaaS apps is essential for reporting and measuring impact

Something like what Stripe Sigma (https://stripe.com/en-in/sigma) did but for all major SaaS apps. Would you want to use something like that?



Would it be nice, of course. However, it makes the assumption that 1) the data is easily understood and 2) that the end result is accurate. IMO, what you'd be solving isn't an accessibility issue, but rather a data cleanliness/data complexity issue. Then I think you run into the problem where the more you abstract the less value you provide. Because if it's too complex, they'll hire a data analyst and if it's too simple, it's useless.

I've heard the "everyone is an analyst" line too, but from business intelligence vendors. We still hire data analysts and developers though.

I do think there's a gap between what SAAS reporting offers and what is needed. For instance, I just led a giant JIRA API data integration that had a significant reporting aspect. Don't even get me started on the Salesforce project.

Baremetrics for XYZ definitely seems like a worthwhile opportunity though.


Are you talking about some sort of:

1) universal plugin to each app that allowed for standard access to raw API data

2) middleware service that you would authorize access to for each of these apps and have the service "know how to join"

3) data lake that allowed one to import all you data into and join together (perhaps with SQL or GQL)


Yes, a mix of 1 & 2 possibly -- am trying not to think too much about the implementation at this stage tbh.

Problem statement remains that every major SaaS app has an API available - while that's good, an API is still an engineer's tool. These apps can't really expose SQL on the go on their apps given a. multi-tenant architectures b. transactional/nosql databases used.

However, an ideal solution does seem like a mix of 1&2 in my mind. An easy to use SDK (?) to make any SaaS app have the capabilities of Stripe Sigma.


Before you commit to SQL, you might want to look at an implementation based on GraphQL, particularly if your audience is slightly less technical.

Bigger set up costs, but a little easier to work with in the end.


We're considering building this as a native feature for our new SQL client: arctype.com.

Let me know if you want to jam on this and discuss :)


Sure would love to chat. Let me know your email/Twitter.


I think the answer is yes look at what salesforce has done with sosl etc very useful and pretty easy to work with.


Thanks - will check out SOSL




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: