Having looked at a lot of these things, here's the criteria I use to evaluate these kinds of services -
1/ How easy is it to get data in?
2/ Is my data safe?
3/ How easy is authoring?
4/ Does it support pivots?
5/ Can I model joins?
6/ Does it support dashboard-wide filters?
7/ Does it support complex time comparisons?
If a service checks those boxes, it's actually ready to get deployed. In the case of Google Data Studio:
1 - Seems standard. 2 - Doesn't use SSL, major red flag. 3 - [EDIT] seems decent, see demo in reply. 4 - Can't tell from the website or examples, red flag. 5 - [EDIT] it cant, see reply 6 - Many examples of this in the gallery. 7 - No examples, red flag.
I'd rate this as pretty mediocre. The people building these products tend to overrate flashiness of dashboards and lose focus on the actual workflow of analysts and those they are helping support. Most of the things I see in the gallery examples are great eye candy but mostly worthless for actually managing a business week-to-week.
Maybe it's different in other companies but I feel that an effective dashboard tells me:
- How I did (last week, trailing 30 days, QTD, YTD)
- ... relative to how I did before (vs comp time period)
- ... relative to my goals
- And a breakout of the drivers of that performance.
Pie charts of demographic breakdowns or measures from some arbitrary date range just aren't useful tools. They're something you look at once and then never revisit.
There are so many alternative ways to achieve the same goal for free or very low cost that unless Google has some secret sauce the product will probably be dead in a few years causing headaches for anyone relying on them.
Obviously if you make thousands and thousands of products, then a couple will eventually have to be dropped. Can you even name me 5 big products they've ended which had a huge impact.
Even if you don’t, I hope Google understands it, and then learns from it. Wanna take a guess at why Microsft kept FoxPro around for so long? One of the reasons was if MS axed it, they’d lose those developers, many forever. Not until they had a good transition story that utilized MS products could they dare discontinue it. And to this day some will still say MS bought Fox Software for the tech and killed FoxPro...despite it being kept around for over ten years.
Point being, MS did the right thing by FoxPro in my eyes, and it is still used as an example of MS discontinuing products. Google has most definitely not done right by some products, so perhaps that will help you understand why people keep bringing it up.
Still they axed Silverlight only a few years after they launched it. Lots and lots of enterprisey stuff was built with it and then all that stuff was suddenly unsupported.
First cut from just checking Wikipedia. Number of google services: 117. Number of discontinued google services: 43.
So what's that.. about 40%?
Still a lot, though.
This is by far the most pervasive comment on HN regarding Google ("Oh, they launched something, don't use it because they will retire it soon. Look what happened to Reader!"). Its also completely misleading when applied to Google Cloud Platform which has been around for almost 9 years now without deprecating any service.
Disclaimer: I work for Google Cloud Platform
If moving the goalposts is allowed then fair enough, it can be any number you choose.
>This is by far the most pervasive comment on HN regarding Google ("Oh, they launched something, don't use it because they will retire it soon. Look what happened to Reader!").
Yes, hello, reality calling - Google has a PR problem, as a direct result of canceling services. This PR problem is bleeding into their commercial offerings precisely because of their reliance and cultivation of grassroots, technology-minded consumer mind share. "Man, I killed all those bees but for some reason now my honey production has dropped right off. Could there be a connection???"
Someone did not properly cost this into the decision to terminate those services, and it's now biting the organization. It should be.
Sarcasm aside, by narrowing the results the goalposts are indeed shifted.
Google's cloud offerings are generally going to have more robust SLA's attached to them, that include various guarantees wrt. service deprecation, that probably compare with other vendors in the same space.
So I do think "moving goalposts" here is probably right. Comparing Reader, Buzz or Google Code to GCP offerings probably is an apples to oranges comparison. In other words, not very useful.
That doesn't feel like the HN ethos. People working for any company should feel comfortable posting here, without feeling like they are the subject of a witch hunt imho.
Through Google Apps there was once a custom URL shortener service ( Google Short Links) that I used to track and shorten hundreds of affiliate links and perform a manner of A/B testing. They sunsetted that service breaking all those links ( and considerable cashflow)
If you dig deep into Adwords or Doubleclick Knowledge base its fairly common to come across broken links, outdated information.
Was a big fan of the Picasa product as I took the time to tag and categorize a huge library of photos which has been long since pulled into Google Photos ( a superior service, but I liked the option of local management for more private photos)
I am very reliant on Google products and services and have been for about a decade. I generally expect that services will disappear, customer service will be non-existent or consist of a series of clueless email from all across the globe. Spend a half million in Adwords, Partner Agency, previously Engage partner - the staff they push on you on that side mostly seem as if they wanted different jobs and generally provide completely dense suggestions (knowledge of adwords interface but no real business acumen ) Adsense! Was once getting between $100-$400 a day in adsense results. All disappeared over night for as single event of clickfraud (likely from a competitor aiming for exactly that result). They even cancelled the last check they sent.
Generally,I've found not worth adopting a G product without an easy exit option or known replacement service.
Even the core search product has suffered but many wouldn't notice - many advanced operators were dropped or limited, regional selection in advanced options minimized, very precise searches made ineffective due to lack of use.
Just meant to say, boy cancelling Google custom search sucked and then reminded how many otherwise Ive been screwed by my reliance.
You can always categorize it different and claim it’s a different story now.
But that won’t convince any customers.
It's not always an easy decision, but my point is that there's a balance to be hit here. There are definitely time where discontinuing products is the right choice and the time is much better spent working on something else. I'd argue that a company which NEVER discontinues anything no matter how old and unused it is is a company that won't go far.
Because it feels like half of their products DO get discontinued.
And because it doesn't matter if it's half or not. As if 20% would be a better number?
One would be foolish to fully trust any company that shut down 1-2 products they were using.
But it would take a special kind of stupid to trust new products by Google who has shut down over a dozen of their products, including widely promoted ones, from Glass and Wave to Video, Reader and so on.
1. Apple (Aperture, "Smart" features from OSX absent from iOS, APIs that force almost complete re-writes, generally culling of PC/corporate friendly products such as Server etc).
2. Facebook, seems like they experiment a lot (which is good) but too often fail to label things as beta. And they change 'features' way too often for my taste (I seldom use their services so far, at least on a personal level).
3. Google (Google Read for news strikes me as the biggest loss personally, my news were never the same afterwards as I never considered putting in so much time carefully selecting RSS feeds etc. if they could go away just like that; eventually I don't even read most news anymore). They have a less-than-stellar record on mobile apps, but then again that's a moving sector where it's hard to be consistent over "LTS" kind of periods.
4. Microsoft, but at this point I have to commend how much long-term support they put into most of their products, sometimes it's even too much to have to put up with legacy features that hinder modern paradigms (particularly on Windows 10, ofc).
5. Amazon, simply because I've yet to encounter a case of product/service being put out of service.
I wonder who agrees/disagrees around here.
- Analog Efex Pro
- Color Efex Pro
- Silver Efex Pro
- HDR Efex Pro
- Sharpener Pro
Right now this product is losing substantial amounts of money, we're talking on the order of tens of millions USD per year. If the net present cost of support exceeds the net present value of the product revenue stream Google has every right to cull it. It's their prerogative as a business.
My prerogative as a business consumer (when I am in that position) makes me wary of being reliant on the whims of a multi-billion dollar company whose primary revenues are generated from ads for my analytics needs.
It's quite possible Google Data Studio is going to be the next big thing for the company, it's just as possible it doesn't get the desired traction and ends up culled from their offerings.
Also, Google can shut down this product without thinking twice. A company that has it as their sole product however, can't.
That's a huge difference.
With Google you usually get a year heads-up before your API disappears
Sure, both scenarios are not ideal, and there are usually other very good reasons to prefer the non-Google solution. But it's not like going with Google is a horrible mistake that would leave you SOL.
That's what I'm responding to.
Of course startups can be far worse. Goes without saying.
from the top of my head:
http://renoise.com/ https://www.gpsoft.com.au/ https://www.ultraedit.com/
Incidentally, they're the best in their respective fields I know of. They don't make products, they don't make product, they make and improve these particular applications. I would LOVE them to expand and make, say, clothing and hair gel with that same intelligent, no-nonsense craftsmanship approach. Not to mention operating systems and search engines. For now I'll dream of Directory Opus for Linux and not even get that.
Feature wise Data Studio has improved a lot since that day, and will continue to do so.
(but just try it out, it's free https://datastudio.google.com/)
They have been saying this is a top requested feature and it is coming for ~2 years. Its a serious limitation.
All Google products use SSL anyway.
I'm relatively new to this field and I found the value of the solution is how it can connect various business workflows instead of just focused on building dashboards.
For example: typical BI tools only focused on Visualizations (most mentioned here only support that and a bit of "Insight"). This alone is not enough because business activities around BI would require forecasting/predictive and planning (and potentially Risk Assessment).
Disclaimer: I work for a business-unit that develop Analytics solution that covered those business activities (https://www.sapanalytics.cloud/product/).
Both are open source and in my opinion can fill about 80% of dashboards and ad hoc analysis for a small to medium company. The other 20 % your company must be rich enough to buy Tableau or rich enough to build its own solution.
Tableau is awesome, but I believe that most of its features are not used soo often. A simpler tool can provide almost the same or more value than Tableau in some cases. Well, a tool that you don't need to pay expensive licenses for each user makes it more democratic, everyone on the organization can visualize and create new analysis.
I totally agree we need something cheap and simple, but it must have some level of "functional" calculations a la excel. Now if users were exposed a pandas dataframe in Superset or Redash or other, that would be extremely cool and would for some cases be much better than Tableau.
Can Metabase do this? I struggled to get a This Year vs Last Year chart without messing with the underlying data.
Is there a good book on Metabase out there?
for instance, for a query like:
select day_of_year, year, count(*) from orders where year >= 2016 group by day_of_year, year
you can create a line visualization that will display the different years as different lines in about 2 or 3 clicks.
AFAIK, there is no book for Metabase. But the documentation and the project is really straightforward.
There's a new generation of these tools that started web-first. They don't have all the bells and whistles but the learning curve is way lower. So far none have truly stand ahead of the rest but there are some promising ones. But because you almost always need a dev-in-the-loop, almost every company needing a tool like this starts with a DIY open source solution.
It depends on the requirements like
1. requirements for having a self host version or acceptance of having cloud offering
2. data connector requirements (not all products will have native connectors to all data sources)
3. data blending requirements (not all products will have data blending capabilities to the same tune)
4. data size requirements (some support data in MBs, some can support data in PB's)
5. Programming knowledge requirements (some may have only non technical users using the product, some may have technical users who need finer control)
6. Visualisation options
7. Various Reproting options (some may have pdf export option, while some others may not have this)
....and the list is really long
There are a large number of parameters in deciding to choose a BI/Reporting solution
One can probably pickup a particular requirement and evaluate the best in class for that requirement. There is no one solution which excels in every single field.
Tableau has been the leader in the market for long.
And the list of offerings is endless
From traditional Tableau, Looker, Sisense, Qlikview
to modern POwerBi, Domo, Tableau online, Data Studio, Bime analytics
From open source solutions like redash.io, airbnb superset, pendaho, jasperreports to closed source solutions like crystal reports...
From simple tools like cyfe , to extremely complex ones
From generic solutions to industry specific solutions like Baremetrics (For stripe analysis), and my own ReportDash (for marketing reports)
We've worked with both small, medium to big tech unicorns. I built the product during my role as data engineer at previous tech startup's employer (we got acquired), so we understood well what are the features desired by SME/startup.
Most of what you are saying is unsubstantiated.
It does use SSL (just like all Google products). Where did you gather that it wasn't using it?
And what do you mean by "complex time comparisons"? There are examples of time series comparisons in the demos.
You're welcome to your opinion.
> 1/ How easy is it to get data in? Seems standard.
It supports a MySQL adapter and a PostgreSQL adapter and then a bunch of Google property adapaters. This is standard.
> 2/ Is my data safe? Doesn't use SSL, major red flag.
The PostgreSQL adapter doesnt use SSL (https://support.google.com/datastudio/answer/7288010?hl=en&r...). This is a showstopper.
> 3/ How easy is authoring? [EDIT] seems decent, see demo in reply.
You can see Felipe's demo here and make your own assessment. https://medium.com/google-cloud/showing-off-the-new-free-goo...
> 4/ Does it support pivots? Can't tell from the website or examples, red flag.
If you find an example I'd love to see it. My expectation at a bare minimum is a table view with a dimension down the left, a dimension across the top, and a measure in the cells. I looked through all the examples and did not see a clear example of a pivot.
> 5/ Can I model joins? [EDIT] it cant, see reply
It can't. https://news.ycombinator.com/item?id=15446616
> 6/ Does it support dashboard-wide filters? Many examples of this in the gallery
Most of the examples in the gallery use this feature.
> 7/ Does it support complex time comparisons? No examples, red flag.
Here's a couple very basic examples of what I'm talking about.
- Plug an TY date range into the dashboard filter. Get a view that shows you TY performance in that range vs LY
- Filter the dashboard using calendar dates, display using fiscal dates.
The rest is written from my personal experiences. If your experiences are different, I'm open to hearing about it. Feel free to email me if you'd prefer.
Yours is a great list. Have saved it instantly! Thank You for that.
That being said, I've used GDS for a few months now, and it has a lot of mid-market potential, since it pulls data from Google Analytics, Adwords, & Youtube out of the box, and, is completely free.
I say "mid-tier" because, for large-scale but relatively simple use-cases (I work on a v. large e-commerce site for a large multi-national, some scale but not the most complex setup I've worked with). While you're "in the ecosystem", you're abstracted from a number of complex issues, like data quality, ingestion jobs, ETL, normalization, etc. HUGE deal for a lot of folks transitioning "into the big leagues".
Commercial-off-the-shelf solutions for a lot of the products companies use are supported by the big ETL startups (Fivetran, Alooma, Stitch Data) to BigQuery.
I think their long term goal is to get these mid-tier folks in and standardized on BigQuery. In the future this may become more competitive with Redshift.
It makes sense. They own the bottom tier. Every team gets started these days by getting good at Google Analytics. Then they outgrow it, and move "out of the ecosystem". Eventually, they may be able to build out an ecosystem around BigQuery for data scientists and data science teams, but filling that gap could make them a ton of money.
UX isn't quite where I would like it, but it's getting better. Some small details make the interface confusing, but that's par the course in this domain.
1/ Getting data in is fine, but you're using BQ. Personal preference that is not my favorite tool.
2/ To my knowledge Google Cloud offers the same guarantees and promises as AWS
3/ Authoring is very easy. Perhaps too easy (I would say this is probably one of the best mid-range tools for data vis. I wouldn't say it wins on a feature-by-feature comparison to PowerBI or Tableau).
4/ Partial (limited extent, but w/o writing SQL)
5/ Partial, but not to the extent you would want in a paid tool in this category. Again, no SQL (this is probably a win for some folks, and a minus for others)
6/ Yes. Although it's more like Filters on a whole dashboard or multiple pages of reports in a project.
7/ I've done custom binning, trailing time periods, pretty much everything I would do on a routine basis for data exploration. However, there's no SQL chart authoring, so I wouldn't be surprised if there's something I'm missing here.
Bottom line, if you're mid-tier it's a great free tool to try, especially if you're just a data scientist looking to do something fancy for a Marketing team with Marketing data, who hasn't really been exposed to rigorous analysis in the past. You could easily do the type of exploration to make good decisions about building an attribution model, for example.
Would meet the needs of 99% of companies I've consulted for, if it wasn't for BQ vs. Redshift.
A lot of my perspective comes from living in a post-GA world. I can see there being value for companies who are reliant on GA and Adwords data for their day-to-day operations - and to be honest, that's probably a lot of people. For someone who isn't as reliant on GA data or already has it in a data warehouse, I can't quite see the justification to use this specific tool. I'll definitely keep an eye on it.
- We are both ETL + BI, we help you get data into your DW easily (CSV Upload, DB Sync, etc)
- Support for dashboard-wide filters, model joins, pivots
For complex time comparison (last week, QTD, YTD) we have metrics spreadsheets: https://docs.holistics.io/metrics/
With or without SSL, exposing your raw prod data to external services like these is a huge risk. Only ever share filtered and redacted data with external entities.
1. A dedicated read-only schema
2. A dedicated user, with only CONNECT to the read-only schema
3. A unique password
4. A dedicated read-only replica DB
you should be safe against pretty much everything.
I'd actually like to be corrected if I'm wrong - this is how I've built numerous externally-facing services.
Something I should have spelled out - the read-only schema has only the data that the charts need (heavily aggregated views). We basically build with the assumption that the schema will be compromised, but only that one schema.
At a previous job a new system was deployed to help the customer service team. I can't remember the details, but it was backed by this pretty meaty database so we could collect lots of data for later analysis. The etl processes was delayed for some reason so a couple of people were given 'temporary' access to the production db. Lo and behold, a week later the whole thing mysteriously grinds to a halt as an analyst left some huge query running in the background that locked up all the important tables until an admin can in and kicked them off.
Letting people run arbitrary workloads on a system you want to be stable is a bad idea.
Obvious advantage of metabase is that it's all local.
Data Studio has a different set of strengths: It's the quickest way I can get an interactive viz published with 0 infrastructure needed. Just build your dashboard, add some controls, and publish it to your closest connections privately, or publicly to the whole world. It will scale without any resource allocation on your side.
(disclosure: I'm Felipe Hoffa and I work for Google Cloud https://twitter.com/felipehoffa)
(in a parallel thread, someone else mentions how they use re:dash and Data Studio https://news.ycombinator.com/item?id=15446296)
Metabase has a link to their repository in the first page of https://metabase.com/
Just based in this fact I would prefer Metabase.
How do they compare to generate high quality reports?
Agree metabase wins though: apt-get install openjdk-8-jre-headless && curl http://downloads.metabase.com/v0.26.1/metabase.jar && java -jar metabase.jar
Also I'm not sure where you bring the "but it doesn't receive bug fixes" statement from? We have a documented upgrade process  for the images you create using our documented setup guide.
Your front page says of the paid version: "Don't worry about installations, hosting and upgrades. All plans include a 30-day free trial. No credit card required."
My first impression was that it is hard to upgrade. Maybe not true, but it is what is communicated by the site.
If by reports you mean a daily email sent with a table and some graphs attached them both suck. The only tool that I know that excels on this is Reporting Services from Microsoft.
But they are really OK for dashboards and visualizations. Metabase has a pulse feature where you can schedule to send a visualization to email or slack, but not something very fancy.
Frankly it has become table stakes to do that.
I do appreciate your reply though, maybe there are new or differentiated capabilities. Would be interesting to hear.
> I can get an interactive viz published with 0 infrastructure needed. Just build your dashboard, add some controls, and publish it
Gartner report on commercial BI/Analytics might be useful to you
PopSQL – Modern, collaborative SQL editor for your team | https://news.ycombinator.com/item?id=15377339 (Oct 2017, 49 comments)
>ghh: Other solutions in the `write sql and share the graphs with your team`-space: https://news.ycombinator.com/item?id=15377339
Franchise – An Open-Source SQL Notebook | https://news.ycombinator.com/item?id=15303833 (Sep 2017, 63 comments)
Disclaimer: I work for Google, but not on Cloud.
Zoomdata is a good alternative, but pricey and not mature enough I feel.
PowerBI is getting surprisingly good, and if you're a Microsoft or R shop it can be a really nice fit for the price. I've thrown a few 50GB+ CSVs at it. It's basically Excel on steroids. If you're dealing with some data stores (like mongo, elasticsearch) then PowerBI isn't viable natively.
Interesting to note that zoomdata and tableau use spark under the covers iirc.
Its got its share of flaws and issues, but for the price I think it beats everything else on the market - especially for existing Office365 customers.
Maybe it is better to use Python Pandas or R interactively to know more about your data.
In the beginning Power BI started out pretty weak compared to Tableau, but over a couple years they make huge investments and seem to even be ahead in some areas. For example, it used to be Tableau didn’t support 3rd party or custom visualization, although I haven’t checked recently if that’s changed.
So much development effort is being poured into many of these competitors, any useful comparators have to be pretty new.
I didn't find this out until yesterday, I've been out of the US and won't be back until almost November. Google didn't email Google Finance users, they just posted a blurb at the top of the portfolio pages mentioning that if you want to keep your data you should download it.
(the story behind: https://medium.com/@hoffa/the-most-famous-reddit-accounts-c9...)
If you use BigQuery and Tableau, you can hook those together, so there's not a ton of incentive to use Data Studio, too, unless it fits a need for sharing with a team.
I can't speak to other Data Studio uses.
I'm no Tableau fan at all, but the fact that it's a native app removes this doubt at least.
My company recently switched to Google suite of apps (Gmail, Hangouts, Calendar, etc) and soon after was given the heads up that we should expect minimal support/updated to Hangouts since Google is shifting their online chat efforts into Allo and Duo.
As for hangouts, they already announced the replacement for this. Hangouts Chat and Hangouts Meet. Meet is already out there and Chat is in closed beta.
We mostly use self-hosted Redash for day-to-day analytics for ad-hoc analytics questions.
"Sorry, we couldn't find any success stories to match those filters."
Once you need something complicated, you will have to or spend decent amount of time preparing the data or back to coding.
selecting the language at the bottom of the page does nothing for me on iOS Chrome. Didn't try other browsers but guessing it's not iOS chrome specific. note my iOS is set to japanese
There is no way to display individual time series data points, but they are always aggregated, and minimum resolution is one hour.
You can have only 10 time series in a single visualization.
(Of course, I just may not know how to use this...)
2. It's free, Plot.ly is $10,000 per year.
Plotly's pricing information is very confusing I must say.
From the looks, is it something like AWS's QuickSight?
I hope Google realises this is a viscous cycle where people avoid signing up for their services, causing low usage, causing them to abandon it, causing lower usage, causing them to shut it down, causing people to avoid signing up for their services...
Also, this is a B2B product which is treated much more seriously than their standard consumer products.
My company recently switched to Google suite of apps (Gmail, Hangouts, Calendar, etc) and soon after was given the heads up that we should expect minimal support/updates to Hangouts since Google is shifting their online chat efforts into Allo and Duo. We're now regretting that switch.
Even if it no longer works for your usecase, it's not a Google originated service but another startup that was shutdown/acquired/sold which is what I was referring to.
I know of at least one website dedicated to track whether or not Google had shut down products yet. I think the thing that started this meme was when Google shut down Reader.
Personally, I've learned to treat SaaS products as throwaway by default. That is, if I can benefit from it short-term, like in 3 to 6 weeks, I may sign up. But I don't expect it to exist in 3 to 6 years, so I'm not going to invest much into using it.
Their numbers definitely are adversely affected by these "acquisitions"
- A startup, which will likely fail
- Not the core business of the company which created it
I'm not very comfortable with giving Google access to our databases.
I use Metabase, and will continue to use it.
Also in the process of setting up Luigi (created by Spotify team) to control dependencies and the ETL process prior to Metabase.
Feel free to reach out if you have further questions.
Then they move everything to one of the big 3 clouds so they can use their whole suite of services...interesting.