Sisense vs Thoughtspot: A Feature Comparison Matrix

For some reason every time I started at a new place, I was thrown right into the BI evaluation process. Having gone so far down the rabbit hole, here are the tools that really impressed me.

• Holistics (we use this)
• Looker
• Sigma
• Hex

Superset has been a fantastic BI tool in our company as an alternative to Tableau or Looker.

Since it is free and fully open sourced it has been great to have our R+D team extend and embed it within our own product, and also empower the BI team to create custom dashboards to replace default ones for enterprise clients.

Power BI. If you know what your doing with that tool and Azure you shouldn't need a dozen other tools. It is dominating the quadrants. It is cheap. Paginated is now included in Pro.

Dwh, measures in dax, blob storage compatibility, great visualization, can migrate cubes to it, etc.

Tableau is not getting investment, others are either immature, dying, recreating the wheel....

My Org. Is currently transitioning from Tableau to PBI. Having used Tableau for the past 3 years I will say it’s more snappy with throwing measures on a chart to do quick analysis. Also building dashboards appear to look nicer than PBI. But having to create a new sheet for every visual can become a pain with data heavy dashboards.

As I am learning PBI, I feel getting data to join is easier in PBI. Not having to do power query in excel then load into tableau to build a data source. Where it all lives within PBI. Although DAX is intimidating, I am starting to understand the logic behind it.

The biggest weakness in Tableau right now, as I see it, is that it is owned by Salesforce. I have no confidence that they can modernize Tableau.

The biggest weakness in PBI it's its visuals. The GUI is very clunky.

Looker is a BI Platform whereas Tableau (Ive not used PBI) is a Visualization tool.

They serve different purposes imho. Tableau does not allow much flexibility when it comes to ad-hoc reporting. A Tableau best practice is to not build dashboards on datasets with more than a handful (20 or so?) attributes, whereas Looker has the concept of 'Explores' that can contain hundreds of attributes for the end user.

A company that only uses a visualization tool like Tableau really limits who has access to pull reports. People at my company complain so much about Looker and they want Tableau but what they don't realize is that if we had Tableau, they likely would not have a developer license and they would have to give their reporting/dashboarding requirements to a BI Developer to create their reports. For small companies that can really create a bottle neck and the average business end user likely doesn't know enough SQL to explore the data.

Semantic layer of using LookML rocks. Visualization piece not so much. Compared to tableau is lacks a lot.

Sounds a lot like us. Legacy environment:

• multiple BI tools / multiple data warehouses
• QlikView: 340 dashboards, 2k+ daily tasks, ETL in Qlik script
• Moving to: 5Tran, Snowflake, dbt, PowerBI

• Qlik script skillset is highly transferable to snowflake SQL.
• Snowflake is a columnar database, like Qlik, and like Qlik you do not have to worry about traditional database concepts like indexing, record locking, performance, etc. Your Qlik developers might find it difficult to find noticeable differences between Qlik script/qvd and Snowflake SQL/snowflake tables beside the obvious syntax differences.
• MSnowflake has all functionality of Qlik script (we did HEAVY ETL transformations in Qlik), and it adds capability like better data insert/merge, and recursion, etc. Some geospatial functionality that looks very similar to Qlik geo analytics which we also use heavily. Obviously, snowflake adds all the new generation of cloud data warehouse capabilities like ingesting different cloud data formats.

• Ensure your leadership understands that cloud means renting their infrastructure. Cost is the biggest pain point of my approach-but it was the choice of our leadership team. On premise to cloud is a big transformation, make sure you know what that means.
• Separate the data warehouse and BI layers, and build good models. Note: Qlik is less picky than other tools about the model because of its associative data models. PowerBI technically works with any model, but you’ll see its best with a star schema. Also, this separation between data and presentation will allow flexibility in future tool changes.

Late to the party, and someone has already mentioned, but I’d like to vouch for Holistics, Sigma and Omni.

They are all decent BI tools, but if you’ve got heavy ETL needs like you did with QlikView, you’ll probably need to add a dedicated ETL tool into the mix. Something like Alteryx or Azure Data Factory could fill that gap.

First of self serve BI is a myth. Second I assume your IT infrastructure can handle either option.

If you want great visualization, drill downs then Tableau is the way to go. If you are in the google ecosystem and also need some ML stuff then looker. Can also look at the volume licensing structure. I would suggest going with one rather than both. Maintenance costs would eat you alive.

Relatively new to Looker development (6 Mos and LookML certified), and nothing seems to work as well in Looker as it did in my old job in Tableau - except the crazy ass sql that can be written by clicking in a well curated explore. Looker Data Studio Pro could satisfy your Tableau hungry people (not as good as Tableau, but it's got more viz capabilities than Looker). It seems those awesome days of analysts building amazing tools, but also holding the keys to the information are numbered if not already gone. Embrace the future as disappointing as it is.

Looker's big advantage, as it's always been, is its semantic layer. It's easy there to map out hugely advanced and complex datasets in ways that then make visualization easier. It's really built as a nice semantic layer that happens to do visualization.

Tableau is kind of the opposite. Tableau sucks at actual data stuff... Managing the data, joining it, building reusable measures, etc... tableau sucks for that stuff. But Tableau makes pretty pictures.

So you might want to think about what matters to you and your users over time.

Seems pretty straight forward - get rid of Looker and make the few users who are using it change to Tableau.

You may have a lower subscription cost for looker, but you need to hold that recurring savings up against the cost of change for your tableau setup: Report migration, change management / re-training of existing user-base, dual licensing during migration period, setup of integrations, etc.

How far out is your breakeven point then on a pure cost / savings basis?

After that you have to consider that in all likelihood, your report migration will not just be a migration but a report change, in cases where Looker can't produce the same visuals etc. as Looker. That's going to add an extra element of stuffing looker down people's throats to get a worse looking product (report) than they already had.

I hope that break-even point is in the near-future. Then you can of course argue that with Looker being part of the google product suite, you've got some architectural benefits and licensing benefits and maybe some machine learning awesomeness you can make use of, potentially... down the line... if it turns out to be useful...

So without knowing your specifics, you'll need to make a pretty strong "architectural " case in addition to your business case. In addition you can possibly argue for a "process optimization" / "time to delivery" case, by having a single platform within the company for people to build expertise in. Basically saying "we're placing all bets on G-suite including BI, that's the direction, your reports may end up looking a bit worse, but we'll save money and streamline our BI process in the mid-term, and long-term we think it'll be a better product".

Haven't heard of Domo but if the Gartner Magic quadrant is considered a good measure then it's a way behind Micrososft.

https://www.domo.com/learn/report/domo-named-a-challenger-in-2023-gartner-magic-quadrant

I'm always suspicious of vendors that say "contact sales for pricing". Any idea what the cost is for licencing? Looks llike you need to write some SQL to connect to data source.

Domo's Marketing strategy involves getting their salesmen in the room with C-Suite execs and keeping "technical people out of the room until we're talking about implementation".

I'm currently with a company that paid a small fortune for domo and it's almost universally hated by everyone. The analysts have to spend hours trying to get it to work, datasets collapse regularly under the weight of compute. It's absolutely not intuitive, the native MONITORING suite is an app you have to install from the marketplace.

At its best, it's a way to go get an excel export. We're currently trying to transition away from it and one of the lead contenders in the meanwhile is PowerBI.

This helps reinforce what we all know: Salesforce killed Tableau (so far).

Also, Tableau is materially more expensive. As they become closer to parity, more companies are going to choose PBI because BI is not revenue-generating (technically). Half the price is a no-brainer unless your usecase is ONLY supported by Tableau.
<br. Honestly, Microsoft should just spend what it takes to match Tableau for viz specifically. Then charge 15-30% more. They would bury Tableau forever because the lack of visual customization is really the only concrete thing Tableau is still significantly better at.

Tableau calculated fields are way more intuitive and simple than DAX. They cannot accomplish everything DAX can, but almost everything. And for the things it can't, it's almost niche.

Still, both tools are quite good though.

You can write sql queries directly in power bi, in the option to import or direct query, extend the window and you can write there. Use it for import or direct query.

I advise building views anyway for reusability and maintainability outside of Power BI desktop app. Im sure Tableau has similar options.

Take into consideration how your end user interacts with the final product and if the aesthetics of one tool or another increase adoption. Those pretty charts are worthless if no one in the company leverages their insights.