How GrubHub gain its initial restaurant partners and customers

GrubHub has a two sided network. So, both the diners and the restaurants are our customers.  I'll answer both sides here, but the restaurant model influenced the consumer side greatly, so I need to answer that first. 

First the restaurant side:

In 2004, my co-founder, Matt Maloney, walked into Charming Wok and sold them on the idea of a subscription based fee for a neighborhood guide.  The next day I quit my job and started going door to door in a very tight geographic area with the same concept. 

By my second meeting I found that I was getting restaurants to sign up on contingent features, so I'd go home and code them that night and then sell other restaurants those new features the next day.  The business model evolved very quickly and within a few months, we had coupons, phone based ordering, search by address, and most importantly online ordering.  (As an aside, this very closely resembles the lean startup movement, though it wasn't called that at the time. We just didn't have investment capital for the first 3 years, and we called it "scrappy"  )

After about 100 of these signups, we flipped the model on its head and switched to a small fee per order.  By taking the risk out of the signup process for restaurants, we perfectly aligned our interests with our customers and reduced the friction in our sales process... 

...which allowed us to focus on our diners...

As you can probably tell, my co-founder and I are both pretty techy. We figured once we had a low friction way to sign up restaurants, we could approach the consumer side programmatically.

The most obvious things were SEO and SEM.  In particular on the SEO side we adamantly refused to build things that were meant primarily for search engine indexing.  Instead, we focused on the idea that the SEO landing pages should be built with HCI principles and conversion in mind.  It turns out that this kind of unique content, highly converting landing page concept also indexes very well, so that worked great.   To boost this value, we added the menus for all restaurants in our markets, and sorted our paying customers (with their online ordering feature) to the top of the list.

Beyond that we tried a ton of stuff with middling results.  Fridge magnets, Door hangers, direct mail, posters in local businesses, coupon hand outs, flyering, you name it.   What really worked was based on local knowledge: Transit ads. Back in the day, when Clear Channel ran the ads for the Chicago CTA, they were REALLY BAD about taking the ads down after they expired. I noticed an add for a May festival up in December.  So, we got a quote for a 6 month run of ads, and then changed the buy to do 6x the quantity for 1 month.  Those one month ads got left up for a long time.  w00t!  Sadly, this doesn't work anymore, but we loved it at the time.

Caveat emptor: advertising beyond a certain point is very inefficient.   What really worked was that we got massive word of mouth referral because we committed early to a high quality product backed up by great customer service.  WOM continues to be our most effective channel for getting new users.