The best way to reduce bounce rates and design a path of least resistance for members and customers is to offer something they want. While others prioritize member acquisition in order to build an early product test case, Sean Ellis has a slightly different philosophy. Ellis has worked with companies like LogMeIn and Xobni to produce scalable marketing programs. Before tweaking design and funnel optimization, Ellis takes pride in the fact that his successes have been built on establishing a core value to users.

Earlier this week he wrote a great article about premature optimization and the mistakes startup companies make in their rushed efforts to convert visitors into customers or members.

Says Ellis, “While this seems like an important goal from the beginning, it’s not. If the value of your core product is weak, doubling the percentage of users that get there won’t help much. And it will actually hurt you because every unit of effort put into optimization is one less unit that you can put into improving your core product. Products that don’t become a “must have” almost always fail.”

Ellis suggests that startups wait to optimize until at least 40% of those randomly surveyed admit that they’d feel a loss without your product. In addition to this suggestion he urges entrepreneurs to be deliberate in both their testing and their analytics programs. The idea here is that instead of having your development team churn out new wireframes every other week, you should prioritize those elements that are both easy to test and more likely to produce results. Essentially you need to form a hypotheses and avoid overburdening designers. If you don’t take this to heart, your tests are unlikely to yield early results and your team will have little incentive to continue in this process of iteration.

Finally, Ellis urges startup founders to consider their most passionate users before changing things. Testing conversion funnels is great when it yields improvements, but if you’re scaring away existing members and customers it’s counter-productive. Ask your earliest advocates and users to prioritize the best parts of the experience. From here you can learn how to capture that experience in messaging and tweak only those parts of the site that are a perceived issue.