I think the development of Python and Jupyter and other less known things like Vega are much more interesting. Python is today the only "glue code" that puts all of it together, from data to insights.
Other than the expensive part, is it really such a bad thing? I feel like relational databases are a pretty good fit for a wide set of use cases and have a huge amount of tooling.
The only barriers to Salesforce + Tableau adoption I noticed were cross-object JOINs and live vs cached data extracts.
Both issues were remedied by denormalizing the data prior to export. For example, a nightly flattened "view" of Opportunities with key related objects moved into columns.
Mulesoft is well suited to perfect the ETL challenges. Bringing them to the table could be a win for everyone.
In that case you may be interested in Dash (dash.plot.ly). It’s a free and open source library that you can use to create dashboards online with Python only.
We write our back ends with FastAPI[1], which is usually just a wrapper around our ML models. Then serve both Dash and FastAPI with gunicorn. The backend is provided the uvicorn[2] worker class with the gunicorn -k arg[3] to greatly increase the speed as well.
For personal projects you can use this stack in GCP's AppEngine standard environment to basically host your (relatively low traffic) apps for free.
1. https://fastapi.tiangolo.com/ 2. https://www.uvicorn.org/ 3. https://docs.gunicorn.org/en/stable/run.html#commonly-used-a...
The real issue has always been the organizational problems of larger teams and companies as data gets split into multiple silos and needs ETL and cleanup before it's useful. The new abilities we have gained have increased the complexity and scale which can lead to new challenges, but the tools are definitely getting better every day.
Don't forget Teradata.
I found the same thing with MicroStrategy. I spent a lot of time reverse engineering what I could from MicroStrategy jars to expose additional functions in their plugin interface (which is so incomplete it shouldn't be advertised). But the reality is its 20+ year old system with front end updates, can only put so many band aids on it.
I think the only thing keeping MicroStategy alive is its cube functionality and the businesses who have invested to much into it.
We seem to prefer the lite version, which is simplier.
If you look at the examples, you can click a button a go to a dynamic editor which we rather like.[1]
https://vega.github.io/vega-lite/
If JS and web browsers aren't your thing, they have python version called "altair"
[1]https://vega.github.io/editor/#/examples/vega-lite/stacked_b...
It's still really early, but feel free to have a play and create an app. Here is an example app examining using the Prophet forecasting library: https://nstack.com/apps/rdA647Q/
I'd love any feedback, and if you'd like to chat to learn more, reach out to me on leo@nstack.com.
To me, Salesforce looks like a big shared Excel file with a bunch of sheets. Tableau... well I can do the same thing with some scripts or spin up a web server.
To others, this tech is just magical. Pay the money, do the integration and it just works... And clearly people will pay a lot of money for things that "just work".
That's a massive waste of your time and effort. The maintenance costs become massive as well should you choose to create your own system. At the VERY least you should leverage existing free open-source tools such as Metabase or Superset or Dash, or free tools such as Google Data Studio or Mode Analytics, if you're not going to spend cash to get a tool like Periscope Data/Looker/Tableau. I mean this gently, but you likely underestimate the complexity of a reliable reporting/analytics infrastructure. Think about it this way - these tools are either collaborated on by a large open-source talent pool, or are created by teams of dedicated software engineers just as talented as you.
I've worked with quite a few companies in an analytics consulting type role, and your "I can do the same thing with scripts" statement is one I've heard countless times. The long-term maintenance costs and technical debt (and "rigidity cost") of rolling-your-own analytics far outweighs the cost of a true analytics platform.
If you decide to roll-your-own anyway, look at tools like DBT and Airflow to reduce long-term maintenance costs.
Yeah I work at a company in the analytics space and see that all the time. It peaks the curiosity of people who are software developers (yet their core competency at their job is something else). They think its a fun work side-project and go after it. Write some python scripts to do ETL and process the data...make a backend with pg, a web server, then do charts in d3.js.
A year later they have a bunch of nice demos to their bosses but nothing that they can actually use in production because it crashes, there's no UI for interactive queries, no reports for people in the business groups, no user management etc. Then they drop it because they're busy with their actual job. So the cost of that engineers time to do something that didn't work was about $20-30k over a year. While the product they could actually use in production was around the same price.
The auto-chart generation is nice. But what about Tableau makes it more likely to be accurate? Aren't you just as likely to make an error on the SQL in Tableau than if you didn't use Tableau?
If you are good, you aren't "scripting", you are making a rad MVC system.
Salesforce puts all the MVC together in nasty, nasty ways.
If this is something you guys are interested in, I started a company called Retool (https://tryretool.com) that is essentially Excel for developers. Imagine if every Excel cell — instead of being a cell — were instead a React component. So you drag and drop these components around, and you can connect them to any back-end datasource (postgres, APIs, etc.). So you could drag on a table and have it pull data from `select * from users` from postgres, and then drag on a button and have it `POST` the selected row back to your API, in order to ban a particular user. The goal is to let end users build CRUD apps (like Salesforce) around their existing datasources quickly.
If you guys have any feedback... I'd really appreciate it. We're just starting out, and really curious to get any feedback from developers. Thanks!
That made it not viable for me. Building from scratch was way cheaper with Upwork. Anyway, Product was cool.
If I had a dollar for every time someone showed me a wrongheaded graph they were using to make business decisions, I could retire.
Is that fully accurate, though?
Where does that leave data scientists / data analysts? I know SQL very well, and I know python's data stack (numpy, pandas, matplotlib, plotly, seaborn, various stats toolkits). I have a strong understanding of the "programming ecosystem" i.e. concepts, terms, definitions, and so on. I understand (basic) computer architecture, I've used and am familiar with (basic) shell/terminal, and services like Docker/Heroku on the command line, and can certainly use GUI cloud tools for AWS, GCP, etc. I can read and understand code and how systems fit together. I've worked alongside engineers of all types.
But I'm not a software engineer. I don't tell people I "program" because my strongest skill is SQL and generally people do not refer to that as "programming".
Never start a land war in Russia, and never neglect your data infrastructure if data is in any way a key business differentiator/fundamental in your market.
Massive customization, and then you are bound by this broke ass Object model in which to get it all done, between Apex and VisualForce nauseating crap.
I don't want to pay hundreds of thousands of dollars for the right to do write database driven web pages.
I thought that the point of Tableau was to provide a tool that end-users (who can't program) could use to interactively explore their data. That's not something you can replicate with a bunch of scripts.
The issue in practice is almost always getting the data in a workable state so that you can manipulate it easily in Tableau. In my experience in smaller and mid-sized places, Tableau tasks get punted to analytics and data science, because they are needed anyway to get and transform the data in the first place. And these people usually prefer and are capable of using more technical tools than Tableau. I know I would rather use Shiny or Dash.
Maybe that's not a difficult problem in larger corps.
And you should know, that the ratio between those two is something like 1:1,000,000
So, for the most of the world, tech is essentially "magic" to them.
Otherwise, it will be hard to justify this high markup for a tool company.
It will be awesome if Salesforce can adjust their model and make Tableau spit out D3. Their desktop tool is nice for designing, but their server components seem frequently unnecessary for running the visualization. The catch is that creating serverless dynamic visualizations isn’t all that money-making and the cool UI/UX design tool is outside of OSS’ wheelhouse.
SalesForce has been pushing Einstein Analytics recently. I haven't used it, but I do see that moving an organization from Tableau to Einstein has a lot of costs involved so this would be a hard sell in many places. Having them both under one roof means they're able to bring a bunch of people across to their cloud and now that license revenue year over year is theirs with the additional data lock-in.
As someone that really dislikes vendor-forced lock-in and generally dislikes the way SalesForce controls your data, rate limits, maxes, licensing, etc, this move is about even more control up your stack that will seem like a "no brainer" to decisionmakers, which is dangerous. That said, I'm sure it will work well for some organizations.
EDIT: I would also love to see them spit out D3 or other open visual but then they'd be losing control of the secret sauce and the requirement for a license. Not sure there is an incentive to go that route.
Precisely what I thought. It looks like their annual revenue is ~ $1 billion, placing this price around 15x annual revenue.
However, Looker has about $131 million in revenue, so their purchase price was an even higher 20x annual revenue.
My conclusion is that these acquisitions are much less about sales revenue and much more about filling strategic holes in product offerings, and I can only assume it's a sellers market in that area.
At $1B/year in revenue there aren't a lot of companies that can realistically acquire you. At $131MM/year, there are.
But still I agree. Both are quite high.
CRM is a pretty darn good ticker for Salesforce. Why would you change it? It perfectly explains your core business
You make it sound like the ticker symbol justifies a $15.7B price tag...
I think we are many years away from $15B vanity ticket symbols.
Perhaps in 5 years we'll look back at this move as a prime example of the bubble we're currently in.
However, 'Tableau Research' [2] has existed for years and its researchers regularly publish at major academic visualisation conferences like IEEE VIS (InfoVis/VAST) and EuroVis (see [3]).
[1]: https://www.geekwire.com/2018/tableau-acquires-mit-ai-spinof...
Can you clarify?
It would be nice if tableau would just generate static content that could be hosted anywhere.
There’s not a client tool for d3 as nice as tableau. I work with lots of scientists who learned tableau but aren’t really programmers and can’t figure out d3 or other libraries.
It's really easy for business users to point it at a database and get started on their own, exploring the data. It feels like it has a much lower barrier to entry than many other reporting tools.
And once you win the hearts of executives, that's kind of the end of that discussion. It's a really sticky product.
The software team was then tasked with putting the fancy graphs that guy did into our web application with the following constraints:
- the Tableau server can't be exposed on the public internet
- our UI can't indicate that it's using Tableau anywhere (i.e., use an iframe or something)
This turns out to be tricky. Tableau doesn't really like to be embedded in 3rd party applications, it leaks information about itself in a number of places. It requires that every person looking at the graph be a user according to Tableau's definition of user. Synchronizing authentication to Tableau server and workbook authorization gets tricky.
The next task coming up is that they want users to be able to ad-hoc schedule an email to themselves with the fancy graphs attached as a pdf, so we've got that to look forward to.
[0] https://onlinehelp.tableau.com/current/api/js_api/en-us/Java... [1] https://onlinehelp.tableau.com/current/server/en-us/trusted_...
This is the endgame of every OLAP tool. But execs really love this feature.
1) An effective tool for people to explore data (with a relatively low barrier to entry - some training required).
2) An effective dashboard authoring tool (i.e. to make small specialized data reporting apps) which are simple enough to be used by anyone without training. These dashboards typically give some sort of situational-awareness for key performance indicators (KPI) such as sales, inventory, etc and are highly specialized for a specific use or role.
>This change in approach could be equated to going to individual airline websites to check routes, dates, times and fares of flights as opposed to just going to a website like Orbitz or Travelocity – punching in where I want to go and when, and it pulls in a report of all the flights that meet my criteria. I can then narrow those results down by a number of criteria – time of day, number of stops, price, etc. It’s self-service reporting in the truest form.
https://portal.key2act.com/Blog/2017/07/Power-BI-choose-cons...
They also do a good job at seeding it in visible places. Alot of newspapers are using Tableau for infographics, etc.
With all that’s happening we’re definitely looking to pick up the pace, and would love to work with more contributors on the free open source alternative at Meltano (www.meltano.com)
Edit: just wrote a quick post with some open questions I'd like to explore around this deal https://meltano.com/blog/2019/06/10/salesforce-is-acquiring-...
Contrast with https://www.tableau.com/, which has a sample graph and "See it in action" right at beginning.
Would like to see a few simple, visual stories about how one could derive business value from meltano, ideally real-life use-cases but if you dont have those yet just make them up.
In the meantime, one way to see more is to checkout our getting started guide: https://meltano.com/docs/quickstart.html and also our YouTube channel which has weekly "Demo Day" videos sharing our progress: https://www.youtube.com/meltano
Really appreciate you taking a look at what we're up to!
This looks great. Will check it out for sure. Keep up the great work.
I wonder if there are any other open source tools in this space?
Our vision is to glue the steps from ETL to dashboard together in an end-to-end solution. We pick whatever we consider best in class and integrate it. So far, we've got Singer, DBT, Jupyter Notebooks and Apache Airflow and we're using VueJS for both the product UI and our website.
We're also working on a blog post exploration what other acquisition might happen in this space. We're adding suggestions to the spreadsheet as we hear them on Twitter, HN, etc https://meltano.com/blog/2019/06/10/first-looker-and-tableau...
It is still beta but should be ready to go in weeks.
Here's an example of the top 100 stories on Hacker News:
https://ohayo.computer?filename=help-for-hackernews.top.flow...
I'm not so sure, I'd probably be too worried about spiders to focus much.
Would it be possible to chat with you about your past experience at Mattermark? I am trying to answer a few questions (not related to the company, but related to VCs and investing). If so, my HN username @ gmail. Thanks!
It’s still very TBD while we get the product to a place where those opportunities start be more emergent. We definitely expect it to be an open core model though, similar to GitLab.
By remaining self-hosted we avoid the big expense (and risk) of storing users data, and they can pick whatever cloud they want. Our team is 5 core members working at GitLab, and we have about a dozen contributors. So it’s kind of a startup within a late stage “startup” (unicorn).
I've worked with a lot of companies who spend months (if not years) integrating their data into a few disparate systems... The finance team has one system (and underlying data lake), the commerce team another, the marketing team another... If Salesforce thinks they can run the entire underlying data infrastructure in addition to the actual customer-facing functions, then this is a smart play.
Data is an asset and liability - when somebody else has all of yours in their proprietary platform and under the control of their cloud, that is a scary proposition to me.
Fair enough if you are running your infrastructure on open standard tech and common cloud platforms, but not locked away in Salesforce.
This is reminding me of the behavior of other large orgs, like Oracle.
This is to prevent cost overruns and solution capture, where every solution to your company’s problems becomes “give it to X vendor” and then X vendor kills a product line and you’re toast.
Salesforce needs to be careful or else they’ll hit that threshold where companies don’t want to use them because you as a client are too small. Google is facing this problem right now.
So while they might lose customers like you, there is clearly ridiculously large piles of money up for grabs if they diversify their products, rather than remain specialized. And, of course, any sufficiently good specialist is at risk of being acquired by one of these behemoth generalists.
There are also cases where a company will pick multiple vendors in an attempt to de-risk and/or for negotiating tactics. If you are fully dependent on a single vendor, the cost of migrating tends to skyrocket and the negotiating power moves towards the vendor.
https://en.globes.co.il/en/article-salesforce-buys-israeli-c...
This deal makes a lot of sense for Salesforce. They should be (and are) on an acquisition spree.
But if I had stock options (or any kind of locked-up equity) in Salesforce, I'd be worried right now. Someone is going to be left holding the bag.
First, how do you know what all institutional investors are thinking?
Second, Sales and FCF straddle Earnings on the income statement. A focus on EV/Sales suggests that investors are optimistic about growth and ignoring the spending required to get there. A focus on EV/FCF suggests that investors are optimistic about increased efficiency and cutting costs.
Not sure if that's a relative bargain compared to this deal or if it makes the $15.7B look totally unrealistic. As of a few years ago the total revenue of the two firms wasn't that different.
But now, Salesforce now has a bigger war chest to play with.
Looks like BI is the new hotness.
Personally I'm really interested in who if anyone will buy Snowflake.
https://releasenotes.docs.salesforce.com/en-us/spring19/rele...
Tableau feels more polished than PowerBI although BI has pretty much reached feature parity with Tableau.
The problem is the licensing costs. PowerBI is very affordable ($10 a month) while Tableau is not($70 and UP).
Tableau is basically saying, unless you have a corporation paying for the license err SAAS subscription do not bother to use us.
https://www.tableau.com/pricing/individual
There was a free option that was pitifuly unused and crippled to cloud only without any exports.
Then again by focusing only on companies for whom the licensing costs are a drop in a cloudy ocean Tableau brought in a 16B valuation.
From there, tableau becomes a "learn it just to get hired at a company that might use it" tactic, which is a valid job hunting strategy but is not something useful to learn in a data viz class, especially since there are other data viz tools the students can learn and keep using after graduation.
Weren't you using offer of free licenses for teaching environments? Those are full versions, albeit they do not include Server, but that's not really necessary to learn how to make solid BI dashboards.
On the other hand, the IOT team at SF probably loves this. I got to spend a few days with the main engineers discussing their "Thunder" architecture a couple years ago. Under the covers it's awesome (great integration of multiple open source technologies) and the "IOT Cloud" UI isn't bad either but they didn't have an answer for data visualization on the scale they built for.
on a tangent, while getting into SF's IOT product they were talking about how it was so easy business analyst could set it up (famous last words right?). It's pretty easy to make a mistake and create a billion SF cases ( case is like a trouble-ticket in SF's Service product) in the span of "click -> "oops let me undo that" -> click" hah
I was actually a part of that team and left Salesforce pretty recently. I think i can safely state that everyone involved in building the tech for that IOT product was pretty proud of what we built.
That being said, the team that built it 2-3 years back and the current team are fundamentally different. The department was thoroughly axed back around November 2017, and multiple teams, including my old one, were spun out into different other projects. It's a shame really, that project was what lured me into joining a behemoth like Salesforce in the first place.
The move to buy Tableau likely comes from an interest in enhancing cross-compatibility between Salesforce+Tableau, the ability to provide a more robust service offering that can compete with Microsoft (e.g. If you buy our CRM solution, we will give you Tableau for a 25% discount) and a concern that another big player would have come in and taken Tableau.
Someone will come and build the dataset (or known as model) and the BA/DS can start building visualization that make sense to tell story about the company (or to tell the story about something).
It's a whole different market/profession.
Disclaimer: I work for PowerBI
Now if you go into a fortune 500 sales meeting to pitch how much better the analytics stack of your CRM is, they will come back saying that Salesforce is clearly investing to be superior in that category for the long haul.
Bad Database Schema === Bad Time
With Google snapping up Looker ($2.6B) for Google Cloud, Salesforce's much bigger purchase of Tableau is a clear sign that the big guys see buying BI tools is a good way to expand the reach of their offerings into more of the business. We talk to companies every day that have made massive investments in data warehouses, viz tools, and high-paid data scientists, but they still aren't agile enough because they can't tie it all together so their people can find and use the right data when they need it. I read an article once about the failure of self serve BI and the reality is that you just end up creating more sprawl. People need tools to reduce the clutter and sprawl and stop the endless chain of emails trying to figure out what table to look at or query to use or if your source is still updating.
We built our data catalog and analysis hub for exactly this, and it's extremely validating to see the big guys like Salesforce and Google investing in expanding the user base of big data tools and I really hope we can be part of the solution of sorting it all out!
Lets break it down:
Both are over priced
Lookeer, however has LookerML
Tableau obfuscates all code.
Looker has easier bolt-ons to redshift/postgres
Tableau's BI tool-set is weaker than looker, albeit, wider spread (more mature)
So, I think google got a steal and SF is playing catch-up... at a high cost.
Plus SF's sunk costs in eveything is going to make a 15B buy take at least two decades to pay off....
(I don’t know why you are being downvoted though, you’re entitled to an opinion same as me)
Or Tableau can continue to pretend that this isn't a real issue and stonewall customers and partners alike.
I think the actual visualisation part is neat, and better than many competitors, but many of the server-side parts are various levels of disastrous (as is their support), and their "data preparation" tool needs some serious improvements to be borderline useable.
15+ billion seems like a lot to me given how Tableau interacts with customers and partners alike, especially seeing how they are activelly alienating existing enterprise customers, all in favour of new sales, but perhaps something will change for the better here.
We're a Microsoft-heavy shop, and I've been trying to get them to move to Power Bi simply because it's far more fully featured, easier to use if you're familiar with the "Windows way" of working, and has streamlined administration/installation/licensing/configuration in Windows environments.
That said, it baffles me why I have to restart Tableau Server 3 or 4 times during installation, and why I have to restart it for trivial changes more generally. For a piece of software that specifically ships with a cluster controller and full-blown zookeeper, somehow their engineers (or "engineers", as I sometimes get the impression) manage to make things that should be trivially solvable with reloads, partial restarts or spawning new workers (e.g. SSL certificates for the built-in Apache webserver) require a complete restart of the whole node.
edit: Regarding Power BI -- I feel that Tableau Server is (for better or worse) one of the killer features for many enterprise customers, because it means all of your data can remain within your own infrastructure and does not have to rely on external cloud providers. If that is not a requirement in your organisation, Power BI might make sense depending on your overall IT landscape, as well as your users' specific needs. On the other hand, if your organisation requires hosting things yourself, I guess it doesn't matter how miserable the experience is for you as an administrator. That's basically Tableau Server in a nutshell.
Ultimately my take is, small business usually take growth from someone else most of the time ( i.e. new restaurant steal customer from another old customer ), instead of generating brand new growth, while bigger business are actually creating - or - consolidating whole industry for growth in a society-positive way in the long run. ( like uber ).
Still somewhere I would wish I can replicate such method of growth for small businesses.
Big companies play by very different rules than small ones, especially big tech companies right now, and I don’t see consolidation being (broadly) socially positive at all really.
I’m not sure this acquisition makes a lot of sense though. The Tableau fad seems to be fading as people realize that while it’s a nice glorified Excel it doesn’t really solve the data issues most companies have. It seems like a very high price for a product that may be past it’s peak.
That said, we’ll done to the Tableau team for having gone public and then selling the company again at a nice bump.
When we started Graphistry, our view was, "skate to where the puck is going", so while we saw Looker and Periscope burning sales/marketing/dev $$$, the result felt 95% similar to Tabluea Public. Instead, we've focused on figuring out how to harness next gen -- GPUs, how to expose ML and automation, etc. While I still think we are right long-term, and that the cloud co's will be paying top dollar here for the next 5 years.. Short-term, I didn't expect this much market hunger for old-school.
Bravo to all the product PMs!
I wonder how this will integrate with Einstein and other AI products salesforce already has. Pardot is the first to come to mind. They already own Heroku, Mulesoft, and Quip.
Exciting times to be a data geek. Hopefully this adds more money to the AI race.
I think they evaluated very well
Even for people who agree on principle, as a business owner in general I'd be very wary of using software from any 3rd party who's willing to try to influence how I run my own business. It makes me wonder about all of the other companies under that SF umbrella too.
Back in the day corporations paid programmers and consulting companies millions of dollars for the purpose of building out applications with databases upon which to digitize their business processes. This was a great for a couple decades, up until some time in the 90's.
It was at that point that expectations had caught up with digitization and having your battle-axes in accounting/order-entry/supply-chain do their keyboard magic on green-screens just wasn't cutting it. To actually understand what was going on, managers had to ask for specific reports from these people and they were just not in a position to do the reports as well has get their own ever-increasing-shit done. Let alone whatever the hell sales was doing.
For a control-freak in management this just isn't acceptable. They started to find ways to "serve themselves". Eventually someone had the idea to hook the enterprise databases into their own zany spreadsheets. It was amazing. There was such a fountain of knowledge and insight that it made the people who could do that seem like relative geniuses. Almost "intelligent" in their business jobs. Thus, "Business Intelligence" was born.
"Business Intelligence" is the practice of making badly-designed opaque shit from brutal, inscrutable business applications... visible.
Only now are makers of enterprise applications starting to get it (or pretend to get it).
It is a great success for a brilliant computer graphics professor: https://graphics.stanford.edu/~hanrahan/ http://www.graphics.stanford.edu/projects//polaris/
My firm uses Informatica's IICS to output all of the datasets for Tableau to visualize, its not as exciting as the hundreds of frameworks and languages that do the same but it works reliably and I don't have to hire expensive data scientists to get the job done or switch tech every few years to whatever is the flavor of the month.
We take the 80/20 approach that most questions the business asks can be answered with Informatica. The last 20% we save for our developers and the odd data scientist to help us with.
https://www.seattletimes.com/business/technology/leaked-docu...
I am personally focusing on Metatron Discovery, which is an Open-Sourced Big data analytics platform for citizen scientists. Link : https://github.com/metatron-app/metatron-discovery
