Google Docs makes it easy to do market research | Hacker News Reader

Google Docs makes it easy to do market research

plus.google.com

207 pointsjasonkolb41 comments

This is a perfect example of Google's absolute failure at marketing docs/apps.

If this had been rolled out as a feature in Apple Pages, we'd have seen it demonstrated on stage, even if only for a moment, and given an easy to remember name (Pages with "Cloud Complete", Pages with "Net Functions", etc.). There would have been deserved oohs and ahhs. It would have been widely reported on and helped ship units.

Google buries the lede and thus misses another great opportunity to show off a really neat feature.

May be this is intentional on Google's part.

I tried a few samples and the result was not always what the G+ Post suggests.

For example, I tried Nimitz, Independence, expecting the rest of the cells to be filled with a few more US Navy Aircraft Carriers and the result was nowhere close.

Advertising/Marketing a "feature" that does not work as intended is a far worse scenario for Google than the current state.

According to the buzzfeed article, the feature was pulled from a discontinued project called Google Sets. So yeah, sounds like they had a cool feature built that is no longer getting updated and they decided to toss it in to google docs.

http://www.buzzfeed.com/justinesharrock/a-glimpse-into-googl...

I personally think this is awesome.

EDIT: ahh, sorry. Just saw the other comment that points out Google Sets. Anyway, this is really cool, playing around with it now, when I should be working.

This is Google Sets, and I recall Google making a big splash about that, seemingly as a way to respond to Wolfram Alpha, which had just launched.

I think you've got your timeline confused.

http://en.wikipedia.org/wiki/Wolfram_Alpha

> Wolfram Alpha was announced in March 2009 and subsequently released on May 15, 2009.[1]

http://searchenginewatch.com/article/2111106/Rock-the-Vote-A...

> Rise of Google Sets

> Originally launched in 2002, Google Sets was truly an innovation in the keyword research tool landscape.

I don't think this feature is that big of a deal. I can't think of a broad audience use case.

This is based on the now defunct Google Sets, a Google labs experiment. It looks at ordered and unordered lists in HTML to find related items.

Here's the official blog post covering this: http://googlesystem.blogspot.be/2012/11/google-sets-still-av...

Actually, I implemented the feature and it was first launched in 2007 – http://googlesystem.blogspot.com/2007/09/google-spreadsheets...

And Google Operating System isn't an official blog of Google :)

So my guess is that (a) the Google Sets feature is discontinued, much to the dismay of the implementing team, (b) they (you?) see a lifeboat in putting this into Google Docs. It's way better there than deep-sixed entirely, but Google Sets was pretty useful to me as a standalone tool - cleaner UI for the capability.

FYI, I don't work for Google anymore.

They shut down the consumer frontend to Google Sets but there are clearly internal instances still running to power this feature (it may also be used in coming up with related terms in another ad tech product, but I'm not sure).

I was new to the Google Spreadsheets team back in 2007 and autofill was a feature that had not been implemented yet. We thought it would be a cool thing to make it a bit better by adding Sets support, so I worked with Simon Tong—the author of Google Sets—to get it playing nice with autofill. To be clear, I did not write Sets. Simon did.

This feature has been live since 2007. It was not created to be a lifeboat for Sets and the frontend getting shut down is the result of (my guess) low user traffic.

Google Sets was one of the best paid-search marketing tools ever.

Thank you for helping keep it alive in some form!

It looks like a great feature! Do you know why was it discontinued?

Google Labs sets indeed. This was one of the most amazing labs ever and I used it all the time.

It's good to know I can use it directly inside Google Docs ; )

This doesn't work for me. I tried putting "chianti" and "merlot" in just like the little video, and after dragging across the cells I just got "chianti" and "merlot" repeating.

ok - my bad. you have to hold down the ctrl/command key when you are dragging.

Only appears to work on Windows. Haven't found any way to trigger it on OSX.

Edit: it's alt on the Mac

Wow, I'd never have thought to try that. I do remember using Google Sets but did not know it made it into Google Drive.

I wonder what other neat features are hidden in there.

I have to use the option key, FWIW. ctrl and command don't work.

My first thought was: will this work with Borges' "Celestial Emporium of Benevolent Knowledge" taxonomy? It (mostly) does.

http://en.wikipedia.org/wiki/Celestial_Emporium_of_Benevolen...

you meant:

http://en.wikipedia.org/wiki/Celestial_Emporium_of_Benevolen...

This works great. I am currently looking to get a small DB of related keywords and just beginning to think about how to get them. This is awesome. It needs two cells to get some context. One cell does not work and should be expected.

On a slightly related note, what do people usually do if they want a list of all say, manufacturers of Sunglasses? I usually am way behind the web server in C and recently dabbling into the JS and python world.

I can only speak about Germany, but there still are tons of catalogs, registers and databases (both in print and online), listing companies that supply specific goods. I am not sure about the US equivalent of Handelskammern or Chambre de Commerce.

Here in the US, a place like hoovers.com would be a good starting point for that sort of thing. There's probably also something like the "Sunglass Manufacturer's Industry Association" which probably maintains such a list. There are also government databases of various sorts, and things like the Chamber of Commerce, etc.

I used the Sunglass manufacturers as an example. Did not mean that literally. I was asking for a more general approach to get lists of related things, exactly like Google sets. Did not know it existed while it did.

Right, I was also just using that as an example. You could substitute the "Sunglass Manufacturer's Industry Association" for the "Mechanical Pencil Industry Association" or the "Nuclear Fuel Rod Manufacturing Association" or whatever. :-)

In either case, the point about Hoovers, various government databases, etc., still stands.

This is pretty awesome - I would use it for very quick brainstorming to get the ball rolling since it's not extremely accurate, but still useful for ten words!

For more keyword suggestions I like ubersuggest.org a lot.

Thank you for this link, I had never heard of them!

I'm a musician so I immediately started typing in a variety of band names and genres. After a few experiments I started to figure out ideal associations. Cool way to discover/rediscover related bands...

I wonder how well the terms correlate with regards to search traffic. I often build up a list of search terms in the google adwords tool and see the traffic estimates for them and related keywords for evaluating niche markets.

Cool trick!

I thought maybe this would be a good way to generate a list of fifty states (without having to find a random page that has a straight-up list) and so I started out by typing:

   Alabama
   Delaware
   Connecticut
   Iowa

And dragged down, however, it only found me roughly 20 more states. I guess there must be a limitation of the query interface...i.e. "find related items, limit 20", rather than: "This user is looking for all 50 state names, so autofill"

edit: And for the hell of it, I entered three of George Carlin's seven dirty words and it only returned a fourth (and a few other non-canonical-dirty-words)

This is off topic but I absolutely love George Carlin. Very smart guy and well before his time. Just went back and listened to his seven words bit, still just as hilarious.

I also noticed that it inserts some empty rows in between. So if it appears to not to show results anymore, just drag it even more, it'll show the rest of them with more empty rows in between .

I got all 50 states when I tried it. Make sure you use lowercase or it'll duplicate the ones you entered.

Is there an app or API out there that does something similar... yet...?

these sets are "interesting" to say the least. I tried it out putting "blonde, brunette" figuring it would give me hair colors.... I figured wrong.

Previously discussed at http://news.ycombinator.com/item?id=4764169

Awesome.. would save a good amount of time.. :D

With the accuracy of data questionable and the universe of data undefined, at best this is entertaining :-)

W/ Jeff Dean cameo!

Should be useful when playing family feud. ;)

Reply on news.ycombinator.com