User-13135008672097594132's answer covers the basic issue very well. I'd add: watch Gordon Ramsey's shows religiously to understand what's really important in running a good restaurant, and things like Jamie Oliver's show.

My instinct is that this is an area where tech can add disruptive value, but you won't get the answers by talking to techies on Quora. You need to be steeped in food culture, the twists and turns of things like the locavore/organic movement etc.

Your competitive advantage isn't going to come from the tech or photos or people. It's going to come from the market positioning and something nobody has yet noticed, having to do with FOOD.

Look at other food-related startups and you'll notice that they don't focus on superficial crap like games or points. The successful ones solve a core problem relating to the dining experience. Open Table makes reservations easier. Yelp organizes good reviews. Groupon and LivingSocial give you the deals to make experimental outings cheaper. These are as close as you can get to core with electronic tech, since unfortunately the Web can't yet transmit smells and tastes. But that's why there's an opportunity. If "reservations" is the closest to the core we can get to, there's a lot of room to improve things.

Remember, you are dealing with one of the most fundamental human drives here. Everybody loves food. It is more fundamental than even sex on Maslow's pyramid. The only thing that is more fundamental is physical security.

NOBODY will use a service around food that doesn't actually make the core eating and drinking experience better.

This means your first problem is to either learn the food market very well, OR find a partner who knows tons about it and find a competitive advantage that has to do with FOOD, not pictures or status updates or social networking.

Off the top of my head, here's a service I'd pay to use: an Amazon-like recommendation system that says "if you liked dish X at restaurant A, you will like dish Y at restaurant B" and the recommendations are reliably high quality.

I have no idea how to solve this problem, but I do know that naive reviews/ratings at dish-level don't work because very few people have good enough taste buds or have sampled enough variety to calibrate their tongues that they can put accurate data into the system.

For example, as an Indian, I obviously love Indian food, and the samosa and dosa are my litmus-test menu items. However, samosa and dosa quality is all over the damn place. It's a crapshoot. And what's more most Yelp reviewers seem to have no idea what makes a good samosa or dosa. I think I do, but on the other hand, I can't really tell a good mac 'n cheese apart from a great one. My tastebuds for that aren't as refined. Even for food I know well, my ratings would probably be much worse than foodies with more refined tastes, but yet my 4 stars would be weighted the same as theirs.

What's more both are somewhat temperamental dishes and quality varies in
the same restaurant. I would LOVE a system that accurately ranks samosa
quality down to batches ("this afternoon's samosa batch at X is
awesome" as a tweet).

I can, will and have driven dozens of miles out of my way to find a good
samosa or dosa. And I am not even much of a crazed foodie. I've known
people who've driven hundreds of miles for a single dish.

Naive crowdsourcing simply doesn't work here. This is an local cuisine expertise, experience and tongue sensitivity issue.

On the other hand the "restaurant critic" traditional model is also unworkable because a few experienced gourmands with cultivated tastes cannot possible sample every item on every menu during every dinner/lunch service every day.

The problem is ripe for disruption. This simply isn't as simple a ratings problem as books on Amazon or answers on Quora. The data is fundamentally harder to measure, and the attributes of a good reviewer are more obscure (you can tell an intelligent book reviewer from the quality of the book review writing; but a beautifully written food review tells you nothing about the quality of that person's tongue).

I don't know why you specifically want to solve a reservations problem, but here's a way to bring food quality into that process. Suppose I've made a reservation at restaurant X because I've heard the samosa quality averages 8.7/10. But on the day-of, very current reviews tell me that today's quality is more like 6. So I cancel my reservation, but it so happens that the chana masala quality that day is through the roof, and somebody else has a watch alert set for that, so when my reservation opens up, they grab it. A food quality stock market with micro-level visibility, real-time reservation management and even stocks/options trading maybe (perhaps I can sell my reservation to another diner or swap mine for his). But again, the basic problem is the data quality coming into the system, and crunching that data intelligently. Maybe seafood is better on Mondays, and samosas are predictably better in the afternoon.

Or as my friend wisely says, "follow the information." To some extent your question sounds like you are a) following the noise and b) are simply tool-happy.

I am not saying you should solve this particular problem (reliable food ratings down to menu item and individual service), but the point is, this is the SORT of problem you should attack. Game points, pictures and social games are meaningless noise unless they help solve a fundamental problem like this. Restaurants are about eating and drinking first and everything else (even relationships and company) a distant second. If you don't solve a problem related to food, you're creating useless dreck technology.