Hacker News new | past | comments | ask | show | jobs | submit login
HN:Help Pivot Our Startup -- Eat.ly (leveragingideas.com)
22 points by shuleatt on Sept 15, 2010 | hide | past | favorite | 32 comments

The idea was born based on the observation that most people can’t remember what they had for lunch two days ago. However, stopping to take a photo (and optionally sharing it publicly) adds both accountability and an element of fun.

A bunch of my friends will occasionally post pictures of their meals on facebook -- either because they're eating out somewhere particularly awesome or because they've cooked something themselves which they want everybody to see. I can't imagine anyone wanting to photograph everything they eat just for the sake of keeping track of it, though.

Definitely need an iPhone/Android/whatever app. Email is too much effort.

Also, a facebook app. Once food photos start showing up in my newsfeed with "Posted from eat.ly" under them, I might start checking it out.

I'd de-emphasise the "health" and "calorie counting" side of things and play up the "food porn" side of things. I don't want everybody to know about the boring healthy salad I had for lunch today, but I do for some reason want everybody to know about the delicious foie gras risotto I had at l'Atelier de Joel Robuchon the other night (as proven by the fact that I just found a way to brag about it in this very comment. nom nom nom)

Maybe provide a way to rate things by delicious-looking-ness? (You might want to think of a better word than delicious-looking-ness.) People could compete to have the most delicious-looking food on the entire site (like hotornot but without quite as much potential for ego shattering).

>Maybe provide a way to rate things by delicious-looking-ness? (You might want to think of a better word than delicious-looking-ness.) People could compete to have the most delicious-looking food on the entire site (like hotornot but without quite as much potential for ego shattering).

The app wasn't entirely appealing to me until I saw this. If you did this, I'd use it. (not sure if there are other players in this space)

Agree with the iPhone app. We do have a basic Facebook integration that allows users to post to their feed. Also agree with deemphasizing the calorie/health aspects, though overtime, I still believe the stronger value prop for an app like this is to become the mint.com of eating rather than another food porn site. But, yes, in the near term more emphasis on social, food porn, etc is prob the way to go

There's already a lot of calorie-counting apps out there if that's what somebody is interested in. dailyburn.com, for instance, is pretty good. It's also a lot of work to replicate one of these since you'll need to enter a huge database of possible foods.

The problem with them is that counting calories is too damn much work, so adding an extra step, where you take a photo of your food and then count the calories in it, doesn't seem like a huge gain. On the other hand, I suppose if you had photos of everything you'd eaten you could more easily count the calories only once every couple of days.

Still, I think you'll need to decide between "food porn" and "healthy eating" rather than try both, since they're pretty much opposed.

First off, have you measured the time it takes to load your site? It seems awfully slow to me (maybe it's the HN posting that's slowing it down) and could be a reason people are not spending more time on the site.

Before you change your vision, have you considered reducing the friction to use the site? Having to email photos and put in the name of the restaurant is cumbersome compared to opening a phone app and taking a picture that automatically gets posted to the site with your location from the phone's gps.

What sets this site apart from the other food photo based sites?

We're definitely aware that site is uuuber slow. Because we're 100% bootstrapped (and not generating revenue) we recently scaled down from Engine Yard to a small instance on Slicehost. We figured we'd cut our costs while deciding what to do. EY was about $240/mth but 10x faster

You're focusing too much on what your users aren't doing. Find out what they are doing and retarget the app based on that.

Precisely. If you could so easily change your users' behavior to suit your needs you would simply get them to pay you money for sitting at home playing Halo:Reach. People pay money to companies that either solve their problems or entertain them.

excellent point

Allow me to offer you an idea:

When I go to Yelp to search for a restaurant, the filters I have available are type of restaurant, proximity, rating and price. My next step is to either sort by those filters, or individually click a restaurant. Looking at an individual restaurant, my options for consideration are photos of the restaurant, consisting of either interior/exterior or images of the food, and user reviews. The user reviews consist of reviews about the service, atmosphere and food, but I have to chug through each review to get this type of specific information. And with a task of choosing a restaurant and having to repeat these sets of actions is immensely tedious.

What's missing?

A way to discover restaurants near me that I can discover by their actual dishes. The user flow would be this -> I'm in the mood for a burger. Show me restaurants that serve burgers. I get a range of places from diners to more high-end to delis and more, but instead if getting a list of joints, I get a sortable grid of user's burger photos from places near me. Then, all I need to do is say, that burger is the one I want and, barring other misc info such as price or needing reservations and what not, I will go there and get the burger that I saw with my own eyes.

The advantage you will have over Yelp is that While Yelp provides a way to discover restaurants by very general filters, you will provide a way to directly discover restaurants by the actual individual dishes.

As a side note, think about when you go to a restaurant and get a menu that has pictures. It gives so much more reference to choosing a dish, because the mystery is taken out of the equation. Your service would do exactly this. Remove the mystery of what you're going to a restaurant for.

Also, while you can see some food items on Yelp in the individual restaurant pages, #1 they come with no context and #2 I can only access these photos on these individual pages which is tedious.

There you have it.

Just want to add that of course, you'd be able to integrate food ratings, friending, favoriting, etc etc etc all that standard social stuff, but IMO, in order for this to be successful, you should consider boiling it down to the absolute simplest of functions and tasks to minimize user friction.

So with the above idea, I would recommend the user need only simply 3 things: 1) photo 2) restaurant 3) whether they liked it or not and an optional 4) description.

If it's an iPhone app then you automatically pull geolocation, but if it's through email, I'm not sure how this would be done and you'd have to explore what you can do.

With regards to encouraging users to contribute photos: Have you thought about creating a few short blog posts about good ways to photograph food? Things like lighting, angles, how to try to take a decent picture using a smartphone, etc?

Food is pretty hard to photograph well, we tend to have overly high standards for what a good food photo looks like because of all the food porn around these days. Just like regular porn, the real thing is going to look a little different. If your users hate the pictures they're getting when they take pictures of their food, they're never going to upload them. They'll probably give up completely after a couple of attempts.

Well version 1 was really not meant to be about the quality of photos (i.e. food porn), but if we do move in this direction it's a great suggestion

Ah, my point was just that if the users don't like the photos they take, they're not going to upload them.

To reinforce this point, we did not have core competency to build an iPhone app and wanted to keep the project bootstrapped. Given the feedback here it seems an app would in fact make a big difference.

Ideas, some already mentioned:

- Android/iPhone app indeed seems like a must, maybe not even so much for ease (tho that would help), but for branding of Eat.ly as a legit product in the consumer's eyes.

- I wouldn't be into the health aspect personally, but I would be into the food porn aspect. It seems too complicated guessing calories etc. BUT maybe you can have two "modes". Turn on "Food Porn" and that is what you get, switch to "Health Porn" and that's what you get. I probably would peek at both.

- This is probably not doable, but if you could integrate menus (like menupages) and calorie counts from restaurants, people could take the picture, pick the food (like on foursquare you pick the venue) and post. Food numbers would come along.

- Single player mode would work better in health mode than food porn mode. You could have numbers to strive for, but then why would you post pictures really? You could just post food.

- Multi player mode could work for sexiest looking food, or for "I'd Eat That!" points. Maybe unlock a badge swarm style.

- Not to mix day job with side project, but it could be pretty sweet to integrate the Eat.ly app (if there will be one) with a "post to foursquare" option so that if you check in your foursquare status can have an eat.ly link to a picture.

As many others have said, iPhone app definitely. This will make the site more casual.

For the "pivot" i'd perhaps suggest trying to get people to document recipes with multiple pictures, this lets you target the more hard core user. Theres an excellent blog that i love reading that gives a good example of what im describing: http://smittenkitchen.com/

I like the recipe ideas, there's not a lot of good recipe content out there, and photos are key

Emailed Eric something akin to the "Hunch for food" idea a while back.

How I roughly imagined it:

Scrap the calorie count. Unless you're at a fast food restaurant where it's posted on the menu, people really are bad at guestimating these things.

Change the health ranking to a taste ranking, and have it be for specific dishes at specific restaurants. While Foodspotting is sort of playing in that space, they are just asking people to submit pics of stuff they like, sans rating. While it's great to know what people like, I also want to know what to avoid. And it's much harder to sort through comments than it is to glance at numerical ratings.

Use case: I go to Corner Bistro for the first time. I do a search for Corner Bistro on Eatly. I see that Eric has had the cheeseburger and that Mike has had the chili and fries. With ratings added, I see Eric gave the burger a 9, but Mike gave the chili a 5, so I'm going with the burger.

In the same vein, using tags and descriptions, aggregate photos and average scores for menu items at restaurants. So I look up Corner Bistro, and I'm presented with a photo (the top ranked photo?) and average taste rank for each item on the menu. It's the Waffle House of Menupages [and I just created a new vertical].

And, yes, eventually it could suggest restaurants and dishes based on my activity and that of friends.

Obviously need to make the entry process as frictionless as possible. Few different thoughts there:

1) Expand current email capabilities, a la Posterous. Use special characters to note the restaurant name, add tags, rate the item, etc. 2) Automate and outsource: map EXIF data to restaurants (imagine you guys could find a database like that somewhere...), superusers for aggregating and tagging, etc. 3) And has been mentioned, iPhone app that makes it all pretty and easy

Man, I saw you guys present at the Quantified Self Meetup in NY (either late last year or early this year) and I thought, man these guys have a great idea, and are waaaay ahead of my "track your daily activities" project. Sorry it didn't take off!

Awesome! That was great fun night. We've tried to hold true to our roots with quantified self/personal tracking. However, most feedback (here and from friends) suggests we pivot away from this

There's something really strange about personal tracking - it seems like such a good idea, full of great potential for people, and yet the same problem happens over and over again - people start using it, fall off and never pick up.

What about a leaderboard or rankings for popular dishes. Users "vote" by taking pics of delicious food. Restaurants would be more likely to promote the site - "Eat.ly votes our cheesesteak best on the East Coast".

You could also run contests for users - each week a different dish. Submit the most beautiful salad, win a prize.

Finally, promoting the site through restaurant weeks might work out well. There are 120+ restaurants participating in Philly's restaurant week right now. Best photo wins another meal.

This is a great idea. The problem with leaderboards and points is a network effect that is not currently there yet. Its a chicken or the egg problem that exists, and we tried to get something out that didn't require mass participation but rather had a strong single player game model. Now that we have some scale (tiny!) this may be a good thing to try

Why wouldn't you have an iPhone app? It would make sending the pictures much easier. It's a small revenue stream, but it is one.

We didn't have an app initially due to resources and an early belief that we could be device agnostic by leveraging email

Also, why not pivot into the diet/accountability market? It seems a bit skeezy, but people might pay for it.

It's possible. One thing we looked at was working with personal trainers, or programs like weight watchers. Another variation was to use mechanical turk to 'guesstimate' calorie counts and charge as a premium service

What is the objective of the pivot? Cash? Users? Pictures? Something else?

Sounds to me like they're startup enthusiasts, and want eat.ly to get bigger and better without sacrificing the stability they have right now, with their full-time jobs.

In other comments, people are saying the iPhone app is important. I think they already know this, and what would be more interesting to them is any shortcuts they could take to build their iPhone app.

I think they ought to find an iPhone developer who has developed an app similar to what would be needed for eat.ly, and is in a similar position where they are enthusiastic about startups but not doing it full-time.

As far as shortcuts I'd recommend doing the "app" as a jqtouch view of their rails? app. That should work for the vast majority of users and avoid the whole outlay for the app store. They could also look at phonegap. It's far better to figure out down the line that you need a full blown app than to figure out that your app could have been done in HTML. iPhone is a pretty big learning curve.

End of the day, probably users and a self-sustaining business model. This was/is a side project among friends so we're not expecting anything too crazy

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact