My biggest strategy is to budget, and monitor the results. For my SAAS products I budget a minimum of 25% of gross profits into re-marketing.
Sometimes this 25% is paid to affiliates as a recurring commission. In other scenarios, we may place a google adwords campaign, allocating $500 for a set of keywords. If that campaign yields $1000 per month in subscription revenue then we would take 25% ($250) per month and continue to grow the same campaign. The only catch is that sometimes we "borrow" money from one winning campaign to try out another one.
We tried this with facebook ads a while ago. We budgeted $250 for a CPM campaign. For our $250 dollars we got a return of 1 subscriber at $20 per month. So, that gave us a reinvestment budget of $5.00 per month for facebook ads. Based on the results it was a failure. Facebook was just not a good fit for our product.
Our number one result has been aligning ourselves with publishers in the form of our affiliate program. If the publisher is not likely to sign up as our affiliate, or is already in bed in an ad network such as adbrite or google, then we place target ads that we track carefully in buckets. The buckets that do well continually re-invest.
I once read a blog from Bob Parsons about advertising that nailed the concept on the head. Good advertising works instantly. Even advertising tailored at just getting your name out there or building your brand should yield near instant results. If advertising doesnt give you a return on what you spent, then its time to move on. There are too many highly comissioned ad executives which will tell you that advertising takes time, and patience, what they are really saying is that if you continue to throw good money after bad then they will enjoy being your butt buddy and raping you.
Tracking what you spend on advertising is a full time job. You really should plan for tracking visits as you build your application. If you use GA (google analytics), make sure that you take advantage of url strings that track medium, source, campaign, and even split testing (although split testing is highly overrated and tends to be done incorrectly, whereas it should be done on focus groups with detailed feedback rather than just looking at results).
Building your SAAS to track where each subscription comes from, you are only a few db queries away to know where to spend your money. Monitor, graph, and pay close attention to this data. Its very sexy when you can see that continued reinvestment gives you logistical growth. Also keep in mind saturation, spending $500 per day on adwords may work, while spending $15000 may not yet you as high as a return. Break it down into buckets, and detail it down it as often as possible.
I run interviewstreet.com - a SaaS based service on screening programmers and I did a few things that helped
Consider the mediums that your typical buyer gravitates to. Launch a blog. Comment on other people's blogs. Do podcasts. Offer webinars. Join LinkedIn, Facebook, Yahoo and Google groups and participate. Create targeted Facebook pages. Launch a twitter campaign.
Neil Patel at quicksprout.com opines quite well on how to get links to your website.
Write articles (or having a pro write them for you) that demonstrate your expertise in your field and then distribute them around the web. This is what my company does. You can check my profile if you want to see more information about that.