Hacker News Reader

After some of the product shutdowns by Google I feel like there's a huge risk building your visualizations on this product.

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.

Replace 'Google' with the name of any of the other startups (or larger companies) in this space and it makes just as much sense.

Not really. Proven trackrecords matter.

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.

Startups fold or get bought and shutdown their product all the time.

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.

> (or larger companies) in this space

That's what I'm responding to.

Of course startups can be far worse. Goes without saying.

According to the above poster, Google has killed 27% of their products. Way more than 27% of startups don't make it.

> (or larger companies) in this space

That's what I'm responding to.

Of course startups can be far worse. Goes without saying.

You mentioned companies with only one product, aka startups.

I don't think this necessarily follows. There are established companies with long track records who specialise in one area.

> companies with only one product, aka startups.

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.

It feels like that's only part of the story though. What's key is which products get killed, and very simply it's anything that doesn't directly make money or indirectly make money by locking you in to Google's core offering. That's perfectly fine, and wholly understandable, but it means products like this data studio that don't directly relate to ads, publishing or mapping are likely to get killed in the future. When people suggest not using it for fear it'll disappear one day they're making what appears to be, to me at least, a rational decision.

So you're saying this should be treated just like a risky product from a no-name startup? That's a pretty sad place to have your product valued when you're in the position of a large company.

I don't understand how people keep bringing up Google product shutdowns as if literally half their products get discontinued. Yes, there are a couple, but that's because they have so many more products than any other company. I'm sure if you looked at it percentage wise, they're actually probably on the lower end compared to most other companies.

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.

People also bring it up for products like this, which is related tangentially to Ads or google cloud offerings. Google is not going to shut something like this down that enhances their revenue

>Google is not going to shut something like this down that enhances their revenue

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.

>there are a couple, but that's because they have so many more products than any other company. I'm sure if you looked at it percentage wise, they're actually probably on the lower end compared to most other companies.

First cut from just checking Wikipedia. Number of google services: 117. Number of discontinued google services: 43.

So what's that.. about 40%?

If 117 is the number of currently active google services, that's 43 / (43 + 117), so about 27%.

Still a lot, though.

Ah thanks, missed that!

I'd say 0, since eyeballing the list almost none of them seem to be B2B type products.

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

>I'd say 0, since eyeballing the list almost none of them seem to be B2B type products.

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.

He's not moving the goalposts. This thread is about a b2b product, which he claims are unlikely to be shut down. If you're trying to extrapolate about a b2b product from past experience about free software, you should justify that.

Specifically it is a dashboard creator b2b product. I bet the google shutdown rate on that is pretty good!

Sarcasm aside, by narrowing the results the goalposts are indeed shifted.

Is it a shift or a correction?

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.

This comes across as a somewhat hysterical response to a fairly measured comment ("Yes, hello, reality calling").

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.

I think your calling that response hysterical is itself a bit hysterical.

Google Maps Engine was B2B example, so your "0" is incorrect

Google Search Appliances, too.

Fair point but neither of these fall under GCP, which GDS does.

That was the point about "shifting goalposts" that was mentioned above.

You can always categorize it different and claim it’s a different story now.

But that won’t convince any customers.

And Google Site Search

Used gCSE as a primary add-on service across a few hundred sites. The twilighting of the product was a major hiccup in our business.

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.

And Google Checkout

And Google ruined one of the most useful assistant features. They may have well shut it down:

https://arstechnica.com/gadgets/2017/04/google-ruins-the-ass...

That's something that irks me endlessly :-(

I've stopped recommending the Home the people because of this. It may seem minor, but the shopping list -> Keep was the simplest, most useful feature that Home had, and Google managed to ruin it.

I am pretty sure the appliance is not discontinued.

It is: https://en.wikipedia.org/wiki/Google_Search_Appliance

A better metric would be average lifetime of a product, otherwise comparing companies with different ages is meaningless. Of course, you have to do something to account for products that haven't yet been killed.

I specifically said products that made a big impact. Look at those 43 services on that list, and tell yourself, would you have engineers waste precious time maintaining that service which hardly get used, or would you rather have them spend said time on developing on newer better tool that have a much bigger impact?

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.

I don't understand how people keep bringing up Google product shutdowns as if literally half their products get discontinued

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.

A company I used to work for still uses FoxPro to this day. A friend of mine at another company is still using VB6 for their core app (though they are planning a rewrite in .NET/C#.) Give Microsoft a hand - their discontinued wares still manage to work on the latest Windows releases. Sure, Google’s offerings are online, so there’s an ongoing cost associated with keeping the lights on, but if Google wants companies to trust them, it’s going to have to stop discontinuing products when they get bored with them.

> 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.

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.

> Can you even name me 5 big products they've ended which had a huge impact.

- Analog Efex Pro

- Color Efex Pro

- Silver Efex Pro

- Viveza

- HDR Efex Pro

- Sharpener Pro

- Dfine

These were big products? Hasn't Photoshop advanced to the point that these plugins are no longer needed?

>I don't understand how people keep bringing up Google product shutdowns as if literally half their products get discontinued.

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.

Again, you say "feels" like, and I agree it's a common trope that shows up in almost every single Google thread, which is why I explicitly asked for people to name real products that are gone. I'm actually curious how many people can even remember or name, because to me, it has become more of a myth than reality.

Indeed, I see them as middle of the road. In my very anecdotal perception, by order of 'less reliable' to 'most reliable' in terms of product continuity, of the big 5 my list goes as follows:

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.

It only takes one product you relied on heavily, to make sure you never get locked in like that again.

Google typically only shuts down Services that already have very low user counts relative to their bigger offerings. Reader was a huge exception, though its closing was more of a migration (to Google+) than a shutdown and it was done under the pretense of moving people to G+.

Do we have something similar from AWS?

Exactly what I thought. Why would you choose to build anything sizeable on something that could be shut down?

An authoring demo I wrote the day Data Studio went public:

- https://medium.com/google-cloud/showing-off-the-new-free-goo...

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/)

> 5/ Can I model joins?

They have been saying this is a top requested feature and it is coming for ~2 years. Its a serious limitation.

I don’t get why no ssl? You almost have to go out of your way to not support it.

I'm not sure where the author gets that. This demo uses SSL: https://datastudio.google.com/reporting/0B_U5RNpwhcE6SF85TEN...

All Google products use SSL anyway.

If I’ve understood, they’re talking about the psql database connections, not the website URL.

Unless your data is pretty small, you'll be wanting it in bigquery not postgres anyway.

Pardon the curtness but "Your data isn't big enough for security" is the stupidest justification I've heard yet.

BI/Analytics is a tricky area with many sub-areas.

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/).

Dashboards Grafana and Kibana enable a lot of clients to self host and create the exact dashboards they need, per customer, per employee etc. Storing stats and data in elasticsearch, influxdb etc etc.

What is the best-in-class of these things?

From a capabilities perspective, Tableau and Power BI have everything and the kitchen sink. But they tend to be $$$$ because of licensing costs, paying for extras, and because you need lots and lots of training for devs, analysts, and end users. Oh, and practically Windows only.

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.

I use and recommend metabase and redash. Metabase is more polished and easier for business folks. But redash supports more data sources (hive for instance).

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.

https://www.metabase.com https://redash.io

> 7/ Does it support complex time comparisons? No examples, red flag.

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?

yes it is, as long as you can represent it in SQL.

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.

+1 for Metabase. I'm currently running a pilot project with it at work, and everyone seems to love it so far. The UI is simple enough that even "common" (as in technophobic) employees feel comfortable using it!

Not sure what kind of organizations you have seen, but Metabase/Redash would max cover 20-30% of the use and Tableau 80%. It depends a lot on the type of data you have, but anything beyond sql will not work with Metabase or Redash.

I think it depends of the organization. Most companies still use something similar to a DW/DL that speaks SQL, so Metabase/Redash still cover this.

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.

But it really doesn't. People will try to do simple calculations as they are used to do in Excel and they will quickly find out they can't do much beyond a simple pivot. Everybody I have see using Tableau is using calculations and a lot of them even LOD calculations. Do this exercise, check what business people really do with Metabase and Resash, I have a feeling they use them to download data into Excel...

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.

If you want something that's affordable, yet can serve the full need of what open-source tools can't, do check us out - https://www.holistics.io

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.

Sorry for having to say the "cliche" answer "It depends"...But it really "depends"

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)

"I'd rate this as pretty mediocre."

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.

> Most of what you are saying is unsubstantiated.

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.

As someone who has been watching Data studio for long, I can vouch that most of the observations here is true.

Currently this connector does not support SSL. Be careful with the data you send.

LOL!

I happen to build a competing service.

Yours is a great list. Have saved it instantly! Thank You for that.

Joins are not supported, although its been on the todo since its ever went public beta.

I've setup a few (more or less standard) data warehouses in Redshift. Prefer SQL for data exploration.

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.

This is really insightful. Thank you for the thoughtful reply.

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.

The direct and easy integration with BigQuery seems like a potential "killer app" for Data Studio.

As I read through your list, I feel like our product (https://holistics.io) can serve you well on these fronts

- 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/