When Mitch Spolan, a LivingSocial senior vice president, told a panel at Internet Week in New York that the daily deals site was launching an instant $1 lunch deal in Manhattan today, I knew I had to try it. After all, a buck for lunch was so much better than the $8.50 or so I often spend on a Midtown, midday, takeout meal.

Excited about actually having plans for lunch away from my desk, I logged onto the site at 9:28 a.m. to snag a deal from one of the 48 participating restaurants. Two of them were already sold out of their dining spots, and I settled on Latitude Bar & Grill, where I planned to take three of my colleagues. My $1 certificate promised us $20 worth of food.

I checked LivingSocial again at 11 a.m. and now 32 of the restaurants had stopped taking new diners. About that time, the Washington, D.C.-based daily deals site boasted on its Twitter feed that 10,000 lunch deals had been sold. We could hardly wait for lunchtime to try out #dollarlunchday

At 1 p.m., we walked through Times Square on our way to Latitude's 8th Avenue location, looking forward to sliders and an experience. Predictably, we found a line and were told by the bouncer that it would be about a 20- to 40-minute wait. The people waiting with us didn’t seem to mind, after all, getting $20 of food for $1 was still a good deal. But of course, they hadn't eaten anything yet.

Checking in with other deal redeemers, results were mixed. "Made a reservation so no line for me at Rasette!!," said @iamevie. @LukeKingma wasn't so lucky: "A disastrous #DollarLunchDay @ChockFullONuts. Food took 40 mins, chicken tenders/grilled cheese were soggy." Added @DrewShannon, "line near my office nearly around the corner. Not sure if I want to wait that long to save $9 on a salad. Maybe next time."

As for us, we started seeing hungry people leave Latitude, complaining about the service. The LivingSocial rep who was on hand to address any technical issues or customer-service problems (mostly related to coupon redemption) was apologetic, but she also claimed to be new to the company.

After about 20 minutes, we we were finally seated. We had a few minutes to look at the menu before our waitress arrived with some bad news: The kitchen wasn't taking any more food orders. The cooks were so backed up that food was coming out an hour after it was ordered. Lunch is over, she said. The only thing our deal was still good on was drinks.

Now, we expected the long wait to be seated, and even the long wait to eat. But we certainly didn’t expect a denial of service. “Are you normally not this busy?” I asked our server?

“Actually,” she said, “we don’t serve lunch.” The restaurant normally opens at 3 p.m., “and it’s always dead until happy hour,” she explained. “We did this [experiment] to try and promote the restaurant.”

When she left, those of us at the table started talking about the push-pull element of the deal concept. Surely LivingSocial should have done more to vet the vendors that signed on to do this. It’s clear to see what the restaurant thought it could achieve with this instant promotion, but not that it would go so spectacularly wrong.

Was it a mistake for Latitude to try this kind of promotion? Considering a recent study that said on average, one unhappy customer will publicly tell 18 people on their social networks about a negative experience, the bad publicity the restaurant stands to see from this deal fail probably wasn't worth the effort.

As the waitress came back 20 minutes later to check on us, she announced that the kitchen had reopened. “But honestly, it would still take 30 minutes for you to get your food—and our sliders are not that good,” she said. She then relayed the story of another customer who waited two hours before finally giving up and walking out. “I’m about ready to walk out myself,” she said.

We finished our drinks, got a $5.50 check and left a $10 tip. We stopped by the Big Apple Deli, and each one of us spent $7.50 for lunch that we then took back to the office. The grand total of this $1 lunch experiment—$46.20.

At least we got the chance to be social.