Zomato - The Good, Bad and More

3 comments:
Zomato, the app for foodies, asked for some feedback for their site and mobile app. Here's what I have to say.

I will try not focusing on the business decisions like expanding to multiple cities, offering discounts, table booking etc. User experience it is for now.


Good features:

Following are some good features that are part of the basic product and doesn't need to be talked about much.
  • Updated phone numbers
  • Menu's for almost all places
  • usable geo-tagging
  • good pics
  • timings data - "now open" feature is a good filter option. 


What I really like:


1. Microtagging - Even before I get to Zomato.com the Google search result for Zomato tells me more about the restaurant than any other site. They have used microtagging well to show "price for 2" as well. This one data point is the most important decision factor (I am a marwadi! :P).



2. Representation of features via icons on the listing/search page. Easy to understand icons around veg/non veg., delivery, bar, credit card etc.

3. Truthful about data availability. No SEO greed. For restaurants where Menu is not available the menu tab is not shown at all. A lot of others sites would add that tab for the sake of fetching SEO traffic and then say "Menu not available". The same is also conveyed on the search result page by removing the menu icon.

4. Data granularity in "Price for 2" - When the site says "price for 2" - it clearly mentions "without alcohol" which adds value to information but what is even more commendable is that they mention whether VAT, Service charges are applicable. Given the fact that these extras add around 25% more to your bill, its good to know in advance. Though, they need to consider Service Tax as well here. 
Tip - Add that to any liquor-serving air-conditioned place. 


Remarkable stuff: 

Stuff that may not affect a lot of users but the product designer has put some thought behind it. 


1. Right click on logo gives you an option to download the logo. Good for Media/bloggers.

2. No Google Ads- It's not easy to give away that temptation for any business. Not having those mostly irrelevant Google ads adds to the user experience. Direct banner ads of restaurants makes more sense.

3.  The Delhi homepage has "Upcoming Events" section above "Featured User Review" which is vice-versa for Bangalore. I am guessing there is a good amount of A/B testing happening.


What I don't like (Needs Improvement):


1. User ratings - The current state doesn't add much value. Make multiple categories for the review - Food, Service, Ambience etc.. This helps me decide whether this place is good for a date (ambience) or for treating hungry college friends (food).

Also categorize the ratings with time as "All Time Rating" and "Last 100 days ratings". Considering last years rating with the same weightage as yesterday's rating for Service or Food may not be helpful as the complete staff may have changed by now.

2. Summarize the review - I don't have all the time in the world to read 500 words of gastronomic affairs. Manu, for example, writes long reviews. Can't you tell me the gist of it? I believe there are open source summarization libraries available, a bit of algo refinement might be required but this isn't very tough. Do this and talk about it in social media, it will also help you hire better when college kids see that you are making an "engineering" product and not just another web application.

Also, "what is this restaurant famous for?" or "A must try" in this restaurant should be part of the review as separate fields.

Ofcourse, increasing the number of fields for a review can be a tricky trade off as this will add a friction for reviewer. A nice set of "optional" fields might help. Or may be again do some data engineering and pickout keywords from the detailed textual review, the way Babajob does.

3. Click to call - For some restaurants the number is mentioned on the site with an ext. number. You are adding an extra step for me to reach the restaurant. At times when I only need the number to book, or I have seen the restaurant on Zomato earlier, I prefer going to JustDial for the number.

Why not give a Click-to-call feature? Too costly I guess. It also kind of competes with table booking services and I am not sure if as a business you want to go that way. Given the fact that you are already doing Phantom numbers or ext. number based system, click to call only doubles the variable cost in short term but might give you bargaining power with your vendor as benefit around economies of scale kicks in.

That said, fake calls and verifying numbers before connecting could create another problem to solve. In the current state you are able to capture the call duration and user's number so your benefits are covered but user is left under-served. May be using the Facebook connect and only connecting the mobile number mentioned on FB profile might help. This could work as an incentive to connect FB profile to Zomato and then a hell lot of Social rich features could kick in.

Oh, you need to consider the fact that a lot offices don't allow FB access and alot of your lunch time usage happens from office. Tricky stuff.

4. Option Filters in search page- The current state of filters is not much usable. The filters mis-behave at times by either not allowing to choose 2 mutually exclusive options or not refining options based on current selection.

For example, when I select "Veg restaurant in Bangalore with live music" it does not let me filter out for bar. All options are disabled. The listing says that there are restaurants with and without bar for this filter but wouldn't let me narrow down.

On the other hand when I select "Happy hour restaurant in Bangalore" it still asks me for Bar/No Bar. Shouldn't it consider "Bar Present" by default? Happy hour is a sub feature of Bar.

There are other such cases of options being disabled/enabled without much intelligence.

Also, the pop-up for cuisine, location and cost is irritating. May be a slide ruler for cost and auto suggest text box with proximity search for location might help. This needs some extensive study.

5. Restaurant address on the listing/search result page is incomplete. I am guessing this is to keep more white space in the middle region and reduce clutter. Also, may be to increase page views. In either case, making me click once more to see the address is bad usability.



 May be you should show the locality along with restaurant name on top and then the rest of the address below. Example - Mention  "Pizza Hut, Saket" instead of only "Pizza Hut" and show rest of the address without "Saket" below.
Tip- This is also better for SEO as you would have different anchor tags for same restaurant at different locations.

6. Keyword based search is bad -  Provide auto suggest in keyword search as well. Atleast the restaurant names and "cuisine in locality" kind of pages should be indexed in the auto suggest. This will also help you serve better results in queries like "Italian in Koramangala". This is a standard search term but the current search algo uses "in koramangala" as a search phrase and looks for occurrences in review, which is kind of dumb. Ideally, "in" should be ignored. If auto suggest is used smartly, you may not need to use NLP to serve such queries better.

7. Banner ads do not add any value - Should be standardized, like Facebook. This is to make sure they add some value as well. Basic info like cuisine, cost, location should be mentioned. If an advertiser falls under my search result, they should be given more visibility to make sure the ad is as relevant as possible. I currently see random images in ads, in some cases the restaurant name is also not readable. Info mentioned above could be shown on mouse-over atleast.

8. Navigation - "Events" link is missing in the top menu bar. User has to scroll down every time on homepage to get to events page. The menu bar on Events page shows link to Restaurants page but vice versa is not true. Not sure if there's a business sense behind that.

9. The slide show on the top of Events Homepage (I checked for Bangalore and NCR ) is kind of irritating/confusing. It is probably hard coded to show 2 sets of slides in a series. With only 1 or 2 items put up on the slide show, I currently see both slides of same item in a single view. Look at the following image to understand this.




10. Data credibility - This restaurant is trending at #5 based one spam review (as seen on 14th Oct 2012). You have to make sure any restaurant featured in the trending list of the week has more data/review/rating to justify that. Specially if it is mentioned on the homepage of the city. Trending list is a very authoritative feature, don't play around with it.

Few Features That Will Enhance The Site Further


1. Restaurant Trending list -  Break down the restaurant trending list according to restaurant type.
Example - On Punjabi by Nature 
What I see - "#3 in Bangalore this week"
What you could show me more - "#1 North Indian Restaurant in Bangalore", "#1 in South Bangalore"

This wold surface more restaurant on a popular list and give an incentive to more restaurant owners to talk about this leaderboard on social media or at the restaurant.

2. Breakdown the reviewers Leaderboard further by type of restaurant. Show "Best reviewers in veg food", "best for bars", "best for five star places" etc. Idea is to surface more users and let them talk about it in social media. In the current state, new users would never be able to make it to the top of that list. May be make that list time dependent as well.

3. Facebook connect - No use elaborating on this. There is lot to be done that you might already be thinking of. Tripadvisor could be a reference point on how to make best use of it. Currently everything is hidden behind a signup button. There is more to be done before that.

4. Reviews - Only show selected reviews on the info page of the restaurant. Lot of unique content is good but it will also reduce the keyword weightage from an SEO point of view.
For example - If you look at the reviews of the top users for this restaurant, the restaurant name is not mentioned in most of the reviews. The review is adding about 300-400 extra words to the page without mentioning your target keyword even once. May be you should put a header like "Naman's review for Punjabi By Nature posted on 23rd July". This makes sense for user and adds the keyword as well.

Also, in the current state the Review page has only duplicate content from Info page. Since you are giving a dedicated URL to review page it would help (for SEO) to not repeat the content from Info page.

5. Search Results on a Map - This is something similar to what Commonfloor is attempting now for property search and Ixigo has been doing for quite sometime for Hotel search. Let the user point his current location on a map and say "restaurants within 2 KMs from this point." Although this is a feature particularly for mobile but at times people, specially in office, use the Zomato website and try to figure out if the place is close to office. Looking at textual addresses one by one is not really productive. Costly feature + low usage, shouldn't be a priority but will set you apart from competition. Shouldn't be tough for you as you already have geo-tagged data. 

6. Mascot - You should consider having a mascot. Authoritative stuff like trending list, Foodie Leaderboard etc. coming from a mascot will give you higher brand recall by giving a "personality" to the brand. 

7. Listen to twitter - A good number of users tweet one line reviews of restaurants on twitter. A lot times they mention @zomato in their tweet. These could be directly indexed on that restaurant's page. Allow twitter connectivity to Zomato and you may not need to depend on @zomato being mentioned in the tweet. 


Mobile App:

I don't use mobile much. Used Zomato app only a couple of times. The following details are from the Andorid App. 

What I liked:

1. Search page - The distance from current location and color coded rating is nicely done with good visibility. Mentioning the locality along with restaurant name on the search page itself is useful.

2. Navigation - The buttons are well sized and every link that you want to click is well placed. Unlike other apps, I don't need to click the home button always and start all over again.

This is specially useful when you try using the search result > filters > cuisine type. This action is divided on 3 screens but a collapsed form of all 3 screens is visible when you are typing the cuisine. You are never lost. 

3. Click to call - The ext. number for the restaurant is dialled in single click with a comma already placed with the number. Solves the usability problem mentioned earlier for the website. 

I did not like:


1. Restaurants with events - This data is not covered on the app. The whole events thing is missing from the app.

2. Filters with 0 results - Incase when a filter results in 0 matching results, user should be prompted to change search criteria. Or automatically show the next best thing. Currently, the user finds nothing to click and might end up quitting the app. 

3. "Distance from current location" on info sceen - This data is not provided on the restaurant's info screen. This is specially important when I am shown a random restaurant from the "Instant Recommendation" link. 

Would enhance the app further. 


1. Location Aware - Notifications like "3 Zomato users are in this restaurant right now" gives a lot of credibility to your app. I am guessing this would be a battery sucker due to regular monitoring but may be this feature should be turned on only at certain time of the day when you expect more people to be out for Dinner/Lunch. These are like auto checkins that can judge the popularity of restaurant. Analytics around what is the avg. turn around time per table at a restaurant could be a good sellable data. 

That's about it. All in all a good usable product with a neat focus. 

How I learnt driving...

No comments:

I have been driving daily without assistance for a month now. Driving has been a challenge for me, for the fact that I first tried driving about 10yrs ago in 10th standard and till date i have taken 4 full time driving classes, 3 of which were paid.

But how I learnt driving is pretty interesting.

1. I have my own car, it is an old second hand car, so did not have to worry about accidental loss much. The risk is calculated. All this while I would be worried more about accidents then actually driving.

2. I have a real need. Till now all procrastination wasn't hurting as I would be at home only for few days and would either have my brother drive me or get my dad's driver. Worst case, I would take a rickshaw which was even cheaper. In Bangalore though, I don't have that luxury and rickshaws are mostly costlier than self drive.

3. I knew where to drive. All this time whenever I drove I never knew the directions, I was trying to solve 2 problems at the same time - how to take a turn and where to take a turn. I wouldn't be prepared for the route and would end up passing the turn and then panic.

4. I wasn't driving too often until now. This thing needs practice on real roads. You can't be driving in a park, once a week and expect to learn. You have to turn up everyday for a date with the beast.

5. I was on my own and felt cheated. The last driving lesson I took was a deal that "The instructor would make me perfect in 7 days". What he really meant was, "He would be available only for seven days, whether I turn up for the lessons or not, whether I learn to drive or not." On the 7th day he came to me and said the course is over and his fee is due. I had taken only 3 classes in those 7 days.  I understood the terms only then, felt cheated and  then I decided to give it a try without instructor, just to make up for my sunk cost.

What I have realised is driving a car is so much like starting up. There are as many things in your control and the same number of external dependencies. I read somewhere that driving is not about controlling the steering wheel but about predicting what will happen next on the road. I think the same applies to business as well. One who can predict what will happen next in the market, can control it.

Though I might sound childish but sometimes it amazes me how millions of people use the same road without talking to each other and without hurting each other, yet they all reach their destinations. Discipline can do so much better to everyone.