We started making a very general CI service, and it took 3 months to realize that we should refocus on web-apps only. At that time, we started to get a bit of traction: maybe 10ish customers were using it, and after a month we asked them to pay.
Most said they would, but for many, the service was too slow and it caused their tests to fail. So we spent another 6 weeks rewriting the backend to make Circle incredibly fast (we had been using t1.micros, we changed to huge machines with LXC, and many custom tweaks), and then our first paying customer, Zencoder, paid.
So we only got it slightly wrong once in this startup, but we had both fucked up a startup before, and learned a lot from it (mostly talk to your customers, and make sure what you're building is something they really need and want).
For us, it wasn't marketing/promo tactics at all, it was just to focus on a great product. Our biggest growth has come from word-of-mouth, and our biggest customers (eg Kickstarter, Shopify, a few more I can't name) came in organically. We got a reputation as being best of breed, and spent a lot of time focusing on how to improve lives for customers.
Rather than thinking in terms of tactics/promos, I would focus on making sure that people want your product, and they want it so much they tell their friends (in our case, colleagues at meetups and over twitter). That's long-term sustainable growth. (You certainly do need short-term growth tactics too, but its harder without the great product behind it).