Places/Restaurants Redesign- Feedback & Ideas?
We are starting to plan a series much-needed improvements to the places (restaurants & bars) section of the site. We launched places two years ago and it has grown into a database of more than 50,000 restaurants, thanks to the tireless efforts of Chowhound users who have both added each of those records to the database, and helped work around several bugs such as creation of duplicates. We now want to build on that foundation and improve on it in the following ways:
1. Create a more complete database of places by licensing basic venue data from an third-party, and eliminate the burden on the users of creating every record, and monitoring for duplicates.
2. Make it easier to find places by location, neighborhood, cuisine, price, special features, average rating, etc.
3. Offer more and better organized factual information about the place such as hours of operation, alcohol served, website link, maps, as well as tools like a link to make a reservation.
4. Make it easier to find, dig into, and participate in the wealth of Chowhound conversation about a given place. This will involve an improved presentation of Chowhound threads and excerpts on the places pages, an invitation to participate on that page, as well as automatic linking to place records on a Chowhound thread discussing a given place.
We are eager to hear your feedback on these plans, as well as your ideas for other features you'd like to see. We promise to listen carefully to this feedback, and though we might not be able to incorporate every suggestion in this first release, we will certainly consider them all.
Thanks, we look forward to hearing your thoughts!
1. Good. I only hope it doesn't step over what is there.
2. Good ideas, but what is your thought about implementing that?
3. What type of change to hours of operations?. When Places first started, I hated the free-form hours of operation field. I've come to really appreciate it and now hate the way Yelp does it which is often more difficult and inaccurate. I like that opentable has a similar freeform hour field for restaurants to fill in as appropriate.
Alcohol served is really useful info. I've been noting 'full bar' next to those that have one. I'd lke to see the category broken up more than Yelp to
I dislike combined categories as they don't give quite the accurate picture. A wine bar might not have any beer. A brewpub, similarily, usually won't have wine.
It would be nice to have a reservation link.
4. This is really difficult to do and believe me I've spent lots of time thinking about it. Do you have any mockups on what is being planned?
Thanks for your great feedback. A few responses:
1. the licensed data will not replace any existing records, it will just fill in gaps (i.e. fields that are not complete for existing restaurants, and/or restaurants that are not in the database at all)
2. We will provide detailed search options for users to search by the above fields alone or in combination, so you can narrow to a particular neighborhood, a particular cuisine, and a particular price range, etc.
3.Thanks, useful feedback on the hours & alcohol fields, we'll take that into account. And we will be implementing a reservations link, where available.
4. Yes, this is without a doubt the biggest challenge of the project, and we want to get it right. We welcome your thoughts on this, and we are working on mockups for this, and we will be reviewing them with a small focus group of users to get feedback.
Some really high level things I'd like to see
1. Reservations and Credit Cards
Display "NO". Otherwise it is difficult to know if it is 'no' or 'unknown' without going into edit
2. Like almost every restaurant site, have a table of price categories and lose average price. That is too difficult to deal with and I really tried. Prices change. I'm never going back to change a price.
For most sites, there are four levels. If you do more like Menupages, it is confusing since everyone else has four levels. I've been adding that currently to the atmosphere field using $$$$=expensive, $$$=pricy, $$=moderate, $=inexpensive.
Yes, I know pricey is the preferred spelling.
Except for $$$ the other terms are the most used across the many sites I looked at. It doesn't matter how you calculate that ... entree, whole meal ... it seems that a restaurant will fall into the same category no matter how that is determined.
3. Add a section for Event Facilities ... most important is if there are private, semi-private areas and how many they accomodate. I've been adding that to "what's it like" Here's an example
4. Add the option for more than 4 photos like Yelp
5. Be careful about neighborhoods. Don't combine. "Tenderloin/Civic Center/Hayes Valley" tells me nothing. ONE neighborhood with the option of adding more than one nabe when it really does straddle certain areas. There should still be some freeform option for locations not in the main stream. San Jose downtown is different than San Jose Blossom Valley or San Jose Santana Row
6. Let search display the 'what's it like section again'
I've been adding brief descriptions at the beginning so when people search they can get an idea if this is something they want to look at further. The old search is still out there and this is what I mean.
7. Have ten levels of ratings. Yelp's five stars don't do it. Allow for zero ... there are places that deserve nothing.
8. Have a menu option. Either a link or an upload function. Preferably both. There are lots of places that don't have websites or links to the better known menu sites.
re: CHOW HQ
One more strong opinion for NO ratings. Jim Leff probably said it best here, and I hope he has some cred going to give importance to that opinion.
It's the essence of chowhound to report on a place that may have one or two great items in a sea of mediocrity. No rating can capture that and would serve only to misrepresent the place.
Please, NO ratings. Star ratings, numerical ratings, doesn't matter. Ratings make Yelp, Zagat, etc. what they are (which is not discussion forums), and the absence of ratings helps make Chowhound what it is. Besides, currently reviews go on the Chowhound boards as posts/threads, and basic restaurant facts go in place records. Incorporating ratings into place records invites people to rate restaurants there instead of discussing them on the boards.
I hope Chow HQ will have some internal discussions about this before considering adding a rating feature. I feel it would dilute and distort what Chowhound is.
I find it...cute, in a way. I'm not arguing that ratings aren't useful in some contexts. I just don't think Chowhound is one of those contexts, because it is a discussion forum, intended to provide space for back-and-forth discussions, and I feel if people can simply rate restaurants in places, it potentially dilutes discussions on the boards.
Don't know how Jim Leff would feel about it in this context, but I also don't recall him advocating ratings as a feature here on Chwhound. I'm expressing my own opinion, and I hope, as I said above, that the folks at Chow will consider whether such a thing makes sense in the context of Chowhound before deciding to implement.
re: Caitlin McGrath
Very late to this discussion, sorry.
Absolutely, ratings are a bad idea.
First: Chowhound has always been super diverse...apples vs oranges vs plums vs toboggans vs helicopters. Can't possibly rate a bakery that is horrendous except for one item so great it's worth a 100 mile trip along the same rating scale as a pretty good diner, an inconsistent high end sushi place, and an exemplary italian ice cart. Just wouldn't make sense. Chowhound's value is in going super wide, calibrating for each context, and not painting with the same brush. Ratings are a ditzy intrusion on that, don't work in that setting, don't make sense, and send the wrong message re: what we're about.
Second: ratings imply a frozen, static quality level that has never applied to the ever-dynamic food scene. That's the guidebook and newspaper model, but we're in a new century and in a new media with a new model. Places waver, and Chowhound's value is in tracking those ups and downs IN REAL TIME. Not in a definitive way (e.g. "Mildred's soup was oversalted on Wednesday...quick, drop that place half a star!"). But in a more immediate and scattershot (but, in the end, quite useful) way....that doesn't lend itself to ratings. At all.
Third, and most importantly, Chowhound's value lies in its critical mass of expert users. And this is NOT a ratings crowd. In fact, they're here to recoil from that sort of model, so prevalent everywhere else. If we tinker with the formula by introducing this ditzy, unchowhoundish feature and the place will come to feel less like the place where this smart, large crowd has felt like home. i'm not saying people will leave in droves over ratings. it's just a chipping away of the distinguishing personality and values of the place, making it more like everywhere else. And as less chowhoundish people - the type who really LIKE ratings-based set-ups - squeal with glee, their increased participation on the boards will dilute data quality and repel the more expert users who were so very difficult to recruit in the first place. And entropy only runs in one direction.
So i think this is a really bad idea, fwiw :(
re: Jim Leff
I know I am late to this, but I just read your post on another thread, linking to your post here. If there have to be star ratings, could older ratings fall off the star calculation after a period of time, say 6 months? Or once a certain number of newer ratings had been posted? Things do change, and it seems wrong that a restaurant will continue to have accolades or criticisms it no longer deserves, affecting the star rating.
I agree improvements are needed and I think all these proposed changes sound great.
#1 - with third-party database, this means users will no longer to be able to create a new record at all, is that right? I don't have a problem with it, but I assume that if there is a place missing, we could alert the team through the usual methods?
I guess it would also help eliminate totally useless records where a user has simply inserted a town or city name instead of a complete address, ie
It will also help with consistency (I'm an editor and writer). Right from the beginning, I developed a sort of "style" where all entries on the Ontario board were consistent in the way the address and other information was provided, including punctuation, etc. I soon realized my "style guide" was not consistent with other place hubs but at least it was within the Ontario hub, at least my entries.
#2 Love that it will be made more easily searchable. I have had huge issues with this, especially because I deliberately inserted descriptions or phrases that I knew would prove useful for searches later, only to have it not work.
#3 - I agree heartily with rworange's suggestion about private events. Early on, I saw the need to be able search and find places that offered good private party services. So I started inserting "private functions" but then realized that almost any restaurant can offer private functions of any sort. It's the private rooms that people need definite information about -- how many does a room seat, any amenities, audio-visual equipment, round or square tables, prix fixe, etc etc. So I started flagging with "private room" instead and even the maximum number seated for these rooms, which worked relatively in regards to searchability for a while but no longer.
#4 - with the changes to the database and improvements to accessing related Chowhound conversation, this will greatly help eliminate shills posting their own record or a disgrunted diner ranting about a bad experience in the "What's it like" field rather than on the regular threads.
These are my initial thoughts, but will be mulling it over further.
- FIX THE SEARCH FUNCTION FIRST. At least the previous one allowed you to search by Title and (IIRC) other fields RELIABLY. Sheesh.
- Consider Breaking Up the Place 'address' field into street address, city, state, zip/postal code, country. This should improve the accuracy and maintainability of the database. The region attribute possibly could be calculated automatically but be overrideable (sp?).
- Add a feature to flag and hopefully *Merge* Duplicate Places by street address and city. Not so old topics have links to duplicate Places which makes it much harder to 'drill down' from the Place to relevant topics. BTW thank you for renaming the Add Place dialog to include Find Place - much clearer terminology.
- Add the 'Missing' cuisine types that have been pointed out ad infinitum or change them to Tags (allow multiple types per place). Arguably, some existing ones are subtypes of others and possibly could be removed e.g. taqueria as a subtype of Mexican (use tags instead).
- A long lost request - figure out a way to Search for Places 'Nearby' (an arbitrary street address / intersection OR an existing Place) and display them on a map. IIRC Yelp has something like it.
- A long lost request - FIX THE @#$%&* PHOTO VIEWER! Pictures/menus scanned and uploaded to Places are mostly useless at their current display resolution. It's simple - if Chow HQ can't/won't remove the right hand ad column, view the picture in its own maximized, temporary, scrollable pop up window just like a bazillion other web sites.
- After Places are cleaned up some (heh), PLEASE FIX LISTS. I want my private lists private! I want a way to find and bookmark 'favorite' lists. Lists should be a drop down bar heading next to Places (btw rename Restaurants and Bars back to...).
- Ratings, neigh. CH is not Yelp. Restaurant owners should just get a clue - give them their own FAQ written with simple words and phrases in a large type font.
- Price category - I'd rather Read The Menu :-) for myself. $ in NYC will probably be $$$ in many other cities. People on budgets will see $ as $$ etc. Not a priority feature IMO.
- Neighborhoods - basically a tag. Not very useful for searching IMO.
- Improve and clarify tag searching - see topic http://chowhound.chow.com/topics/615998
In summary, give preference to streamlining and fixing existing important features. Regarding 'new' features, I have an alert button and I know how to use it: http://snipurl.com/alertbutton
Thanks for listening. Seriously!
One of the things I use most is the price category. It is one of the top questions asked on the SF board.
Seems to work both nationally and internationally on Opentable.
I was thinking about this and maybe something similar could be done similar to Opentable where if you click on the word 'price', you get a price legend. Here's a Canadian restaurant just to use an example outside of the US
There could be some price legends in the FAQ and help section. Opentable only uses three levels. Unless you want to get nitty about it using opentable which bases it on the entree price
$$$$ = over 50
$$$= 31 - 50
$$ = under 30
Yep, when i'm in some small town, someplace serving $20 entrees might fall into the $$$ category for the location. Still,using the 80% rule, that about covers most places nationwide. I'f I'm spending under $30 I think of the joint as moderate, the middlle as pricey and unless things are really that sad in NYC, wouldn't entrees over $50 qualify as expensive.
It is the stuff on the borders of these prices that are sometimes hard to determine, but whether a place that serves upper $20 entrees mainly is $$ or $$$ isn't that big a deal in the grand scheme.
And ... there is nothing stopping you from looking at the menu ... if there is one.
Pretty much, never looking at a single menu, people know which restaurants fall in what category in their area.
1. Include "BYO" in the alcohol field values.
2. Please split up the Tri-State suburban boards -- Long Island, New Jersey and Westchester/Hudson Valley/SW Connecticut are three very distinct regions spanning hundreds of miles! -- even if you do implement searches by neighborhood.
3. Have you though about allowing users to rate comments and/or commenters?
Rename it or feature it in some way that people actually know what it is and use it. "Restaurants and Bars" is not only inaccurate (there's a lot more than that!), it doesn't convey what you actually find when you go to that site. I would never have looked under that tab unless I already knew what it was.
Add an "in the neighborhood" feature -- I want to be able to zero in on an area on that map and see all the places in that area. In fact, I originally thought that was what this feature was supposed to do.
A couple of suggestions:
-a "cash only" designation....as many 'hound-worthy spots are cash only
-a "late night" filter (ie closing hours later than 11pm-onwards)
-cuisine is a cool feature - though it might be problematic if it isn't extensible (eg.."Chinese" is not really descriptive enough of a category here where I live)
a useable mapping /search by proximity to address feature so that a person looking for a place in a neighborhood can tap into the site info and find out what the options are nearby
website and menu link
I am in Minneapolis, stuck in the catch-all Midwest. One thing that irks me about Places/Restaurants is that it highlights the geographical bias of this site. For those of us without a home state or city board is there is no way to create a Sticky for any of our cities.
While I applaud your revamp of Places/Restaurants, to me it’s like Scarlett O’Hara when it comes to creating new boards, “I can't think about that right now. If I do, I'll go crazy. I'll think about that tomorrow.”
I know Places/Restaurants is a potential revenue stream so I get prioritizing over new boards. Still, it stings a little.
To make this constructive my suggestion is to have Places/Restaurants not so tied to ‘home’ boards that we in St. Paul-Minneapolis (and every other boardless city) can get a sticky. or any other clear means of restaurant categorization
re: MplsM ary
I totally agree. It is absolutely IMPOSSIBLE to expect someone in, say, Wisconsin to even consider launching discussion about the merits of various bakeries in his/her hometown on a board titled "The Midwest". It'd be like discussing some small town's parking lot issues in the US Senate! This geographical situation was an embarrassment for me when I ran the site, and it continues to embarrass. But I have two thoughts:
1. it's a miracle Chowhound continues to exist at all (and its future existence is not assured; CBS, unlike me and Bob - the site's founders - needs to make a profit). So while there are all sorts of things that ought to be, hey, at least this much is hanging on, and we'd miss it if it were to go away.
2. start local discussion sites for local discussion. I always encouraged this in the past, and it's still a smart idea. It'd be great if there was a practical way for Chowhound to penetrate every last burgh in existence, but, as-is, it's too blunt a tool for that, except major cities. So come back here for general issues, cooking issues, gadget and cookbook issues, and for when traveling. And post momentous local finds here, too. But if you want to discuss the relative merits of linguini served in various Italian restaurants in Fargo or Madison, there are surely locals who'd join, so why not build a discussion group? That sort of thing won't hurt Chowhound; you'll keep coming back here, too, I'm willing to bet! In other words, use this site for what it's good for, and, for everything else, be resourceful!
Note that I'm not anyone "official". Just another hound giving an opinion.
In this post I will concentrate purely on location. The queries I would like to is of the form...
Select * from places where....
type = restaurant / supermarket / market
day = tuesday
time between 16:00 and 19:00
distance < 10 miles / km
locationtype = intersection / postcode / zipcode / lat-long / town
location = 33133-4111
cuisine = korean or vietnamese or cambodian
pricerange between 2 and 7
chowhoundrating > 3
description contains patio ( like %patio% )
distance / price asc /desc
One problem lies in deciding which information about a restaurant should be free-format and which is confined to a restrictive drop down. Ratings and prices are examples mentioned above. Free-format is great for data entry but lousy for searching. After displaying a list I would like one click access to the details or posts concerning them. This includes the expansion of all previously-read posts and being located at the first relevant post in the thread. Ideally, all posts referencing the place should either be highlighted and / or have a suitable search string inserted at the beginning of each relevant post. So instead of a post that reads...
"I agree that Drunk as a Skunk is the best pub in the area."
It would display in pale green...
I agree that Drunk as a Skunk is the best pub in the area."
Possibly Drunk as a Skunk would be dark green. This would allow easy scanning of relevant threads.
And one other thing I would dearly love would be to download a complete board's restaurant data in the form of a spreadsheet.
Since they are starting testing, I'm guessing a lot is already chisled in stone.
I am hoping the time is freeform. I was absolutely against a freeform field when Places was put in place. However, as I used it over the years, there's a lot of reasons it should be freeform. There's other ways of getting to a search on restaurants open at a certain time, but I'm guessing
a - even if there is a static time field, we won't be able to search on it
b - the programming would be outside of what this site does.
Unfortunately, I'm guessing that there is going to be a static time field. That means I will be updating that info less because I'm pretty much cutting and pasting info from websites and it is too time-consuming (and not as accurate) to input time the way Yelp does it. Even yelp doesn't allow searches on open time.
chowhoundrating > 3
Well, as a Chowhound poster, despite being ok with ratings, that would get me insane just on terms of linguistics.
CHOW places is NOT CHOWHOUND.
There is a subtle distinction, IMO, between the two. For all the people against putting ratings on places ... again, Chow Places is not Chowhound.
So my objection would be calling it CHOWHOUND rating. Call it Chow rating. There is a difference.
I feel like the Chow vs. Chowhound diff in Places is semantics. While the Places feature may be called a Chow feature, for all intents and purposes, it is a Chowhound reference tool, because all the data therein is essentially fed by Chowhound users, in the context of Chowhound discussions. Who uses Places? Chowhound users; ergo...
re: Caitlin McGrath
To me the fixed information is chow stuff. What the people think of it is chowhound stuff. But as a system designer I know only too well how difficult it is to design a new system when there is already one in place, one that is loved (or at least people have a vague attachment to) and already understood.
I get the feeling it would be too late to change a concept which must already have passed through most of the design stages. But before it is built I would have liked to see some screen shots of the initial design. By the time you get to beta testing there is usually too large an investment in time, resources and egos to make radical changes.
As an example, let us look at ratings. There are a many of ways we can approach that. Here are two different ways with some of the benefits and problems.
Select a word... Poor, average, good, excellent, whatever
Pros: Easy to code, easy to understand, fixed values in database etc
Cons: Lots of them...
Language and semantic dependant
Max 5 or so values
User feels it is time consuming (not the same as IS time consuming)
Feels old fashioned with no feedback (ie boring)
Alternative: Slider bar with grumpy face on one and and smiling one on the other. Changes colour as we move from left to right. Sort of like:
:-< ------------!------------ :-)
Cons: Not initially intuitive, much more difficult to code, especially in an HTML environment. Depends on user 'feel' for where to drop the slider.
Pros: Language and semantic independent. Granularity (ie you can have 0 to 10, 100 or 1000 if you want.) Gives people the feeling of more control. More enjoyable (maybe).
I must presume somewhere in the corporate dungeons these types of resource / pay-off decisions are being made. It does strike me that to some extent we now have a legacy site in chow.com. With the advantage of 20-20 hindsight they would have done things differently. But it is too late to make radical changes.
Another site will do that.
And I'd love to see a physical data model of the database.
>let us look at ratings. There are a many of ways we can approach that.<
All of which ways ignore the fact/opinion that ratings are a good example of what should not be included in chowhound because they cannot express the essence of a place, cannot describe the place with one or two great items in a sea of mediocrity, and would simply misrepresent may places. Leave ratings to Yelp.
1. Chow places vs. Chowhound places
Fact of the matter is that chowhounds put information in places. And are potential users. Why build something that chowhounds would not want to use?
2. Building yet another restaurant listing
There are tons of websites with restaurant listings and ratings etc. If we want to build yet another one that people will use, it's got to have real utility above and beyond everything out there. Otherwise it's a waste of effort. And a poor investment of time and money. Just because the listing will have links to chowhound discussions etc. doesn't make it better than average.
3. Some concrete ideas of how we could be better and different
(This comes too late, but better than nothing.)
a. allow people to upload pictures of foreign language menus and post translations
b. if we want ratings, have ratings for every single dish - that way we can see how the fine grained ratings are distributed, and one-hit wonders on the menu will show up. By looking at a histogram of the dish ratings [imagine mousing over to one of the bars on the histogram, and a list of dishes pop up, colour coded by course (apps, main, dessert etc) , main ingredient] one would easily see what dishes are "good" and whether the restaurant is "good" as a whole. People could search by the average dish quality, the quality of the fish dishes, the quality of the best dishes etc.
c. cuisines should be an unlimited list of tags, rather than a choice from a drop down menu
d. location should be live and modifiable by the restaurant owner. This is particularly useful in the case of street vendors or food trucks. Parse updates via twitter if necessary.
Do whatever, but do it in a way that is worthwhile and makes places stand out above and beyond what other restaurant listings do.
As far as point b,FoodieBytes breaks down ratings by dish. I have yet to see one dish that received a rating. Personally, I don't think like that and for me I would find it tedious to rate each dish. In the end it wouldn't be any more useful than a overall rating. That great soup yesterday, is swill a week later when the chef changes. Zuni in SF is known for its chicken, yet not one person has rated that.
Then there is the changing nature of what is on the menu. So many restaurants are serve dishes that are seasonal and dependant on the chef's creativity at the moment.
I'm kind of ambivalent as I mentioned on ratings. I ignore them on most sites and read the reviews. However, a new poster on my local board was requesting a rating system and as an example of the usefulness of ratings, someone found this ...
Example plucked from the internet.:
5 of 5 stars
Jun 25, 2008
1925 SW Wanamaker Road, Topeka, KS 66604
Very romantic and relaxing atmosphere. Sinatra-style music all the time.. You can't find a fancier restaurant with these prices. The service is always wonderful. Last time I went, I got the all-you-can-eat soup, salad and bread sticks. I asked the waiter for a to-go container for my full bowl of soup. He gave me an empty to-go container AND another full with more soup.
That post will probably get deleted, but I thought it might be appropriate here.
All right. I'm officially annoyed. I know, who gives a ..... BUT,
Has anyone had any actual contact with the Redesign team or seen the beta?
What's it all about, Chow?
This has probably come up before, but I do hope the next rollout allows users to add or modify place links upon editing. More than once I've neglected to add a place link before posting, and then found that I had no second chance.
Places is a feature that Chow is trying to promote. I wish the site would try harder to make it easy to use.
I only just noticed that they've added the rating system. I've not been part of what has apparently been an ongoing discussion, but...WHY?
Many people have already articulated this sentiment, but it really seems to me that the lack of ratings (and, thus, the emphasis on narrative, qualitative reviews) has always been what set Chowhound apart from Yelp, etc.
In other contexts, I enjoy rating things as much as the next person, but here, I really think it's a step in the wrong direction.