Is Tableau Dead?
mergeyourdata.com
mergeyourdata.com
Needless to say it got nowhere fast and, by the time he gave up and threw us “back” to renewals, the damage had been done.
We elected to do what they thought was unthinkable and we just walked away.
My oh my, it was very, very good day, when I got to tell them to go f themselves.
Feels like software sales has just gotten worse for years now.
Shade of its former self, expensive, and there are plenty of better options. Really seems like the company is in zombie mode until it's eventual slow demise.
It has this feeling of "just works", and documentation is pretty good.
Sure it might not scale but Heroku didn't either.
We chose Render and have been happy with things just chugging along as they were on Heroku.
Their runtime seems to take up quite a bit more RAM, but the IAC is a great addition resulting in 100% consistent environments.
Now I avoid AWS as much as I can. Of course my employers tend to be on AWS so sadly I can’t completely avoid it, but I do my best and would never recommend it anymore.
It is a tool like any other and it is not as simple as DO's but its also a lot cheaper and more powerful once you learn the tooling. The logging on deployment is true but I think the majority/all issues I have seen with that is when the image is not configured correctly and cannot run when testing locally.
I don’t remember what I was charged for, I assumed it was because it was counting the few seconds of running time at the minimum billed time each time I deployed, but I’m unsure. This was in 2020. All I know is that I had a reasonably simple setup, there was an issue where it wouldn’t successfully start up, but logging was flaky at best and I was charged for my attempts.
I’ve been using AWS for years, at various employers and my own last startups, and it’s far more complex than many of its competitors. That’s ok when you need all of its features, but if you need just one or two, then the complexity gets in the way.
Github Action (Build Docker Image) > Push Image to AWS ECR for image storage > Kick off Deploy of ECS.
You can configure your ECS cluster using the console UI or something like Terraform. Once configured its set it and forget it.
Honestly not that much more complicated compared to Heroku.
Just wondering how do you "Kick off deploy of ECS" services? The solution I have currently is using:
aws ecs update-service --cluster={ CLUSTER_NAME } --service={ SERVICE_NAME } --no-cli-pager --force-new-deployment
aws ecs wait services-stable --cluster={ CLUSTER_NAME } --service={ SERVICE_NAME }
From my CI, but this doesn't feel like the best way to do it
I don't think the way you are doing it is the worst way. I am biggest expert here of course. I do mine via a github action, right now it triggers on merges to main but could also be setup independently.
Condensed but I use 1) "build-push-action" - github action to send an image to ECR 2) "aws ecs update-service --cluster <name> --service <service> --force-new-deployment" for the deployment
In a linear graph from "Heroku" to "VPS do-it-yourself":
|------------------------------------------------------------|
Heroku Render Railway NorthFlank Fly DigitalOcean
I use Render for https://www.gamedrop.gg/ and it works really well for me, very hands off.
I saw Fly.io and it doesn't even offer a managed database, not to mention the fact that they have had serious reliability issues.
Heroku is pretty battle tested and mature at this point, though the feature development has stopped, I don't doubt Salesforce is now going to use it simply as a the "cloud" offering for their other products..
It is a shame, however, that the behavior of the sales rep probably made the renewals rep's chart look bad if not worse.
Letting something "die on the vine" is the only way to describe Salesforce's Tableau acquisition.
probably made the sales rep chart for "outreach activity" tick up
and the renewals rep's chart was fired because they didnt hit their target
I can't decide whether this statement is ironic or not. Salesforce is just another big company, with committees and meetings and crisp charts that exist only to show numbers going up, not to actually identify anything as boring as actual operational performance.
Is the fate of commissioned salespeople to be eaten by other commissioned salespeople.
It took some time for me to switch from Excel paradigm to Power BI paradigm, but once it clicked, I've never looked back.
https://news.ycombinator.com/item?id=16477003
"I've now 4X'd my pricing (from 3 figures to 4 figures a month!) and the customers got 10X better, complain less, I work less and earn more"
Imagine for example you're a small shop with your own budget in an org the size of Alphabet with a couple of Java servers that you purchase support to the tune of $5k/year and then they all of a sudden want to charge $1,000,000/month because there are ~200k employees within your larger organization.
It gets even more convoluted when you get into hardware support, where the support itself outside of warranty can be more expensive annually than replacing the hardware and getting N years of support with warranty.
PowerBI are their lunch because it was “free” with the rest of our Microsoft stuff, so everyone could make the shiny dashboards no one uses instead of just a few.
In some cases we just wrote our own d3-based visualizations. It was easier than dealing with the constant licensing headaches Tableau brought.
I worked on so many that eventually i came to hate making them lol, data viz can be extremely subjective and you get into weird paradigms of how much data viz you display vs user actions on that snapshot of data (ex. viewing details, exporting to diff formats, printing the dashboard with no cut-offs, manually refreshing stale data vs auto refresh)
Microsoft like any software company will raise prices on their tools similar to tableau or oracle in the past.
Microsoft uniquely has a large portion of the fortune 100-500 as customers. They have already had that market from office, m365, etc and will grow those things.
Microsoft is relatively early in the licensing lifecycle arc for cloud.
This will happen relative to the maturity of offering, uptake/demand to pricing.
Some of this can already be seen with the Business central pricing.
Again, the above is not to single Microsoft out. Only that their products are relatively young but developing extremely quickly and the execution of the pricing models will be relative to a lot of things like demand, and stock prices.
At the time the incumbent, Qliksense, was lagging badly and there was genuine grass roots support for Tableau as our next gen BI tool. Adoption grew significantly and people were happy...for a while. I remember reaching out the founder/CEO at the time and getting responses back. It was great.
Fast forward today and its different, over time Tableau dissatisfaction grew, I think a lot of it was the fact that our use-cases got more advanced and performance dropped and not being able to understand the complex SQL it was generating didn't help. Also Qliksense got noticeably better UX and importantly was cheaper. Today, MS PowerBI which came from left field, is really the front runner for us.
Parking this situation for a second I think the real issue is that people are getting a bit overwhelmed with dashboards. I have access to near 100 across the three platforms, they more or less do the same thing and are thus somewhat commoditized, I don't feel passionate about any of them as I did back in the day with Tableau. Worse, I have to spend mental bandwidth figuring out which one I need to open to answer my query, then I have to spend time navigating the UX to get my answer so I can move on with the task in hand.
We don't necessary need more dashboards, just faster ways to go from question to reliable answer where information can be pushed rather than pulled, especially if data points start looking anomalous over time.
Note - If your a startup in this field looking to upend the BI space reach out, I have spent a lot of time in this space over the years.
My feeling is that these dashboards were probably more like one-time reports that someone decided to automate. But there's definitely a sense of Tableau fatigue. And I don't know if switching to a different technology like Power BI will make that any better.
I go back to thinking about how the data can be pushed rather than pulled. Whilst BI tools have some solutions I never found anything great so ended up with homebrew scripts. Definitely an opportunity space.
Sometimes you just need to see the data once or month or once a quarter to make a decision. Sometimes, you just need to see the data once! (an ad hoc analysis in Excel)
We should think of the rate at which we need to act on decisions, and it need not always be a frenetic "now, now, now"
You've got daily driver exec and performance dashboards, then you have over time trend dashboards, and finally you have drill down self serve explore of data export tools. Alerting should be separate.
It's expected that most dashboards will not be checked often, trends don't change daily so it's usually not meaningful to look at most metrics daily. Self serve tools are only needed occasionally, but exist as a cost savings on data teams doing low effort high time tasks.
Dashboard tools and spreadsheets are the only good tools for sharing data, I find most reports and spreadsheets are actually just one offs.
Denial: "Tableau will become a tool in the Salesforce toolbox that major players will continue to use."
Anger: "Salesforce executives mentioning Tableau less than Slack or Mulesoft on their public calls ... Meanwhile, Microsoft's Power BI is on a hot streak."
Bargaining: "I believe it'll be a consistent player at large institutions who typically go through complex project and procurement processes."
Depression: "It will no longer be the hot new tool that will be embraced by SMBs. The community will not have the hope and excitement it had in the 2010s."
Acceptance: "The magic is no longer there and that's ok. Nothing lasts forever."
The licensing costs are prohibitive, too. It is almost unaffordable to share Tableau dashboards with your customers and, instead, have to send them static PDFs of your dashboard.
Beyond that, the Tableau infrastructure feels like something from the last century. Data is cached on the server using data extracts rather than following a more modern caching infrastructure.
The caching of Tableau is a huge money safer and huge performance boost if you e.g. use it with BigQuery (cut our BG costs by 80%).
As with every data tool, you need strong data governance, or everyone will define their own metric for revenue and then you have endless discussions about "why the data is wrong" (it isn't).
The biggest benefit to me is every marketing person knows Tableau and with a self explaining table structure (e.g. DBT on top of BigQuery) is self serving.
The only good thing with Looker is their data journeys, which Tableau doesn't have.
The worst (for EU/GDPR at least) feature of Looker is how people can send emails from a query (circumventing corporate email infrastructure, blacklists, data protection etc.). Can you say "Intern sends spam mail to millions of customers because they thought this was a good idea".
Beyond that, it is also a major GDPR headache in Europe. If someone uses their personal account, with elevated access, you suddenly have a data extract on the Tableau Online server that contains personal data, with no way to enforce retention periods.
Google laid off their US-based community support team (https://www.reddit.com/r/Looker/comments/t5jpwk/google_cloud...) and in the time since, those docs links have rotted/disconnected.
Even if 100 US-based community support folks were making $300K in total comp, why would you not spend $30MM to keep your $2.6B acquisition humming along? A bit mystifying.
https://github.com/ankane/blazer
No such luck in my current role, Looker and PowerBI are both in use by different bits of the org and nobody has the ability to delve into the underlying figures.
I like the notebooks feature, the integration of output cells with reports is so so, but it's the best balance between exploratory and reporting tool I've seen.
I found the parameterized code powerful but creating a basic report with lots of filters and a few parameters was very slow.
Here's an example, you do a customers rollup and an orders rollup, but a user asks for average order value of return customers buying product X. In SQL, that's a semi join of order line items, on your orders table and an inner join on daily customers to determine if they're a new customer. But no BI tool I've ever used has semi joins, so when you join in the tool your order values are multiplied and you're back to custom SQL to fix it.
It's that or build a whole new mart to answer this one question.
My current company sells telephony and call-center analytics for MS Teams using our software, and during pre-sales meeting, one of the staple questions is "So we can avoid Power BI entirely?" followed by a minor thanksgiving ceremony and presentation of credit card. I get a feeling it's not too dear to the hearts of its users.
To put things in context, our small claim to fame is anyway getting some aggregate data about lost calls and attempts that is not possible to produce in Power BI directly[1]. So it's likely that they would buy or not buy it even if it was based on Power BI itself.
It takes a lot of trial and error and patience to finally grasp how Power BI handles calculations if you are used to the freedom of Excel.
That being said, once it clicks, it feels like you are a magician, and you can do really advanced analytics on giant tables, which is great.
We pull a bunch of data out of the guts of MS Project 2010 and do a lot of custom analysis on it. It took about two years to go from scratch to a set of reports that served everyone's needs that also required minimal maintenance.
Then we got purchased by a larger company, so now the work is in rolling out a similar set of reports for the sister companies, as well as rolling them up into corporate-wide reports. It's a lot of fun.
Most people just don’t to analyze data and hardly even know how to use Excel, so something like Power BI or Tableau looks like rocket science when really these tools should make you feel like you’re upgrading to power tools from basic hand tools.
PowerBI / Tableau is much more full-featured.
What’s nice is that it’s better than Excel so if you want to replace Excel, PBI is great. But if you want a functional data viz that you’re not embarrassed to put on your web site, I’m not sure PBI works.
The reason I say this is because something that would take me hours to do in Tableau maybe only take a few minutes with Power BI’s mashup language (Power Query) and the Functional Language DAX. The ability to create temporary tables on the fly and utilize variables and the hundreds of other things you can do with DAX to create complex measures, calculated columns, etc. feels empowering. Most people I’ve spoken to who’ve gone away from Tableau to Power BI and used it long enough start to really appreciate it for the vast capabilities.
Additionally, the latest Power BI visual calculations feature released has made the learning curve a lot shorter for those newer to DAX and all our lives much easier.
On the visuals side there has been a lot of entrepreneurship to create custom visuals with the SDK to do a lot of visuals that just weren’t available in Tableau. So many that I have a hard time going back to Tableau because its missing a visual I need.
Lastly, not many talk about the fact the IT loves it in that its got so many capabilities for providing data governance and security through its various features.
Update and one final thought: there is some genius to me in how Microsoft seemed to have created an abstraction that allows the platform to grow independently and greatly on so many fronts - DAX with new functions and features, the Power BI Visuals and their SDK to do more and keep in with user interface changes, and their BI model approach (which now allows GIT versioning with readability changes).
I’ve done the microstrategy, powerbi, tableau dance for about 3 years and I wasn’t impressed.
It’s easy to do easy things and hard to do hard things.
The versioning is quite bad.
The powerBI gateways are complex to maintain, expensive and incomprehensible when they don’t work (update failed? Good luck! Out of quota? Times out? Haha…)
DAX is pretty; but when it’s big, it’s spaghetti. Worse than SQL.
The custom components are frustratingly difficult to get right because they have to be extremely flexible.
…but, sure; it is what it is.
The ecosystem is pretty good.
Lots of documentation on how to do trivial things.
It’s ok.
I would avoid it if I could though; there’s a deep well of suffering to anyone who has to look after the stuff that people “whip together quickly” in PowerBI.
So… I guess it’s fair to say opinions vary.
Maybe it’s better than the others?
…buuut, like it? Nah.
Wouldn’t say love but prefer over the alternatives. I still find the user interfaces could have been designed better if they were not so focused on matching like other office products. I wish it was as nice as it is to wireframe in Figma. The ribbon and other office elements are dated and not as conducive to todays standard where good collaboration tools is helpful to bringing developers and designers together while being efficient.
> The versioning is quite bad.
Agreed, but it’s got better over the years. Especially in the most recent release of Power BI projects. I feel that Microsoft’s strategy is to provide this new format and let independent developers create versioning tools while they take on higher priorities. They have often taken things those developers have created and incorporated them into their products. As with all the product requests, you got to pick your battles - I don’t defend which ones they choose. I’m just happy to finally have GIT versioning.
> DAX is pretty; but when it’s big, it’s spaghetti. Worse than SQL.
I agree 100%. I liked how Tableau had a more SQL feeling syntax. I luckily come from an Excel heavy background so adopting was easier than some. I find that most of the verbose code is due to dealing with row and filter contexts which in DAX can be frustrating. Debugging is difficult… I can go on. With that said, others have offering differing opinions (see: https://www.sqlbi.com/blog/alberto/2020/06/20/7-reasons-dax-...). All I can say is that I feel there is probably could have been a better syntax / functional language then the one created to produce the same functionality and maintain independence.
There are impossible to drill down and understand - "where did this data come from" - how was this calculated.
Similar to other comments, you can version, it's beyond hard to error check. A python or sql script you can at least verify and recreate. I view Power BI as just the new excel - big files full of hidden errors.
At least tableau project files are just xml so you can stick it in git and have ugly merge requests. And rollback to versions if necessary.
The fact that Microsoft doesn’t even see this is a problem is infuriating and makes it hard for any serious data work where you need reproducibility and auditability.
https://powerbi.microsoft.com/en-us/blog/deep-dive-into-powe...
They recently released an update in preview that has a more human readable format:
https://powerbi.microsoft.com/en-us/blog/tmdl-in-power-bi-de...
For comparing semantic models there is ALM Toolkit: http://alm-toolkit.com/
You can also edit BI Models with Tabular Editor which can save objects to storage:
https://tabulareditor.com/downloads
And phi-tools for DevOps: https://pbi.tools/ I’d wait in pbi-tools to start working with Power Bi Developer Mode / Power BI Project files
While what we wanted was something that was SQL-native.
Back in the days we opted for a tool called Chartio which was later engulfed by Atlassian/Jira. Not sure what happened to it after, but it suited our purpose fine enough. Just line-charts and easily editable SQL.
I have some experience with Superset, but it was not really built for customer facing dashboards. Mainly useful for internal use.
Blazer as someone else posted sounds interesting. May check it out.
https://github.com/evidence-dev/evidence
Previous HN discussions:
https://news.ycombinator.com/item?id=35645464 (97 comments)
https://news.ycombinator.com/item?id=28304781 (91 comments)
Yes, Salesforce consumes companies whole. Just like they did with Heroku and others.
People think of Tableau as a dashboard tool, which it is. But it's also a great tool for doing EDA (exploratory data analysis) on multidimensional data. Instead of writing lots of seaborn/ggplot code (and having to remember the syntax), you can just move things around in Tableau in a fraction of the time.
The downside of Tableau is that you have to have a strong relational database way of thinking to use it effectively. You have to think in terms of aggregations, and occasionally in terms of window functions (Table calculations in Tableau). Tableau was (is?) a much more powerful tool than Power BI (at least this was true in 2019 when I last used PowerBI).
Tableau has never been a simple tool by any means. I was a scientist but I used to run Tableau training in my workplace, and I could see how long it took for things to click for most folks. Once you got beyond a few bar charts, things got really complex (partly because you were unlocking a lot of power through relational operations -- in data analysis, there's rarely power without complexity)
Relational thinking is a very powerful way to think about (and execute) computation on large-scale data in a performant way -- it is after all the algebra of dataframes -- but I've come to appreciate that it is beyond most people. Even software engineers struggle to think relationally -- many still think in for-loops.
Also, things that were seemingly easy to do in Excel -- like doing a row and column cell calculation -- were very difficult. You had to think about the corresponding SQL manipulation and then reproduce it in Tableau....
... but I think most people just want to plot nice graphs from their Excel data and show it to their bosses. (which they can do with Tableau, but so can they with Power BI)
Side note: I have a copy of Tableau 2023 but it seems to have reached a feature plateau -- it is barely different from the Tableau of 2020.
Statistical tests are good but they are summary statistics. Often you can get a lot of insight by just looking at the raw data. Tableau lets you see the actual outliers and ask yourself why they are there and drilling down to find out more.
The simple act of looking at the data as-is and asking the questions — before jumping to statistical testing — is an underrated discipline because it seems too simple and unsophisticated. But it yields insights that many statistical techniques will not.
I had the same experience. Before, my company tried several times and failed to implement enterprise dashboards and reporting. I was able to champion Tableau and get it off the ground and even the board of directors peruse the dashboards. I thought I had a huge win. I could not believe the amount of negativity I since received.
That is I believe enterprise reporting is essentially a people problem. The people who looks bad because of the metrics will always blame the tool. They won't be happy until they can manipulate the tool and massage the data until the metric becomes meaningless. In other words, it is convenient to blame the reporting if their report card shows they are failing.
I now use both of these tools below (no affiliation), which you might find interesting.
[1] www.exploratory.io
[2] www.count.co
A year or two really isn't a lot of time (for an individual nor company) to invest in order to build long term skills in powerful tools that side step inflexible (and expensive) platforms.
1) Premium (24/7) support if needed.
2) In-house tableau consultants.
3) Too many production reports have already been written in Tableau.
With that said, I've experienced some frustration with Tableau. Want writeback combability? That'll be thousands (if not tens of thousands) of *dollars per year* in additional license fees.
Want some other feature, same deal.
With that said, the licenses have grown so much in the past few years that our stakeholders have started asking around for analysts to check out competitors and OS tools.
We mostly use Looker Studio + Metabase currently.
Investigating Tableau, Looker, PowerBI, Sisense, Omni currently. PowerBI is great, but BigQuery integration seems to be lagging. Is there a way to get such a tool without paying $50K a year? Everything so far is only a "nice to have" upgrade over our mostly free toolkit. We also care about good geospatial visualizations and embedding reports for users without requiring a login.
https://docs.snowflake.com/en/developer-guide/streamlit/limi...
They're aware that our platform might hit some snags with the more complex use cases but most of their customers have similar use cases that are less complex.