if you’re in a crowded space and need significant publicity/broad user generated buzz initially, makes sense to add sexy features and push the PR side

if you’re in a less crowded space and looking to slowly build momentum/a solid user base, then makes sense to thoroughly Q&A and focus on stability/performance

i think 3 is something that can be layered onto both but should be second priority

When you are in the last week before a final launch adding new features is absolutely the wrong thing to do. It’s tempting - I’m sure they could actually get a new feature done (and A/B testing is actually just a feature really). By trying to add in something new you risk adding instability and you risk just having the new thing be “bolted on” and not make any sense. And if you feel that your product won’t standup in the marketplace without it - then change your launch date and do it right.

But more importantly, if you are chasing that “one last feature” during the last week, you are missing the chance to do all the other stuff you need to do (press, marketing prep, QA, etc). Take the time to signup for the product from a clean browser, watch friends signup and use it for the first time. Fix anything glaring and note all the rest. This is your time to see the product as the world will see it - that one broken “help” link will stop a user - did you find it?

And most important: sleep. The launch isn’t the finish line - it’s just the start.

The utopian world of shooting for both looking good and working good is never really achieved – product designers will always find ways to make their products look better (remember the first iPod?) and engineers will always want to rework the product to make it perform better or have more features (anyone want to go back to iTunes 1.0?). I could concede that given nearly infinite time and resources a product could be made that both looks good and works good, but I have personally never had access to or desired to work in that environment.

So, let’s say that you make a product that looks good. Yet, the product has limited functionality and may not have all of the features to hit your target market perfectly. Good looking products get to market and attract attention for their companies, even if they lack features and performance. If done correctly, say like the Apple MacBook Air, the product still gets good visibility and adoption. Further, because the product looks great, customers can see where the product is headed and usually attracts early adopters.

On the other hand, let’s assume that the product does not look good but works good. If the product is loaded with features and functions and just needs some work to make it look good, then it can also attract the attention of the target market and early adopters – think about early Blackberry products and early versions of RedHat Linux. A product that works good but does not look good gives the customer the ability to justify using the product because of features while waiting for the company to refine their look in subsequent versions (which both Blackberry and RedHat have done).

The death zone for products are those that fail at either looking good or working good. For a product to be successful it has to excel at one of these two objectives – if you are not there with your product, keep working and hold the launch.