"No experiments" was the famous slogan used by the Germany Christian Democratic political party in the 1957 election. It remains one of the most successful election slogans of all times and helped secure a landslide victory for Konrad Adenauer.
Having spent a few years travelling around the world talking to online professionals frustrated with failed projects, budget overruns, unstable websites, intranet downtime, vendors acting cowboy-style it seems to me that you as customers need to stop experimenting.
The solution is clearly not to continue with more iterations and perpetual beta releases. These might work well in a market with speed and constant change, but that's not what's really happening. In fact, very little has changed when it comes to online projects over the past years. The skills, user experience, strategies, planning, goals and technologies are not really changing that fast.
Instead of blindly experimenting and hoping for the best, use the vast knowledge that already exists to find tested solutions to your problems and requirements. You can tap into an endless source of inspiration by talking to your peers willing to share their lessons learned. If you are not able to find a reference for a given user experience, product or intranet application, it might be wise to wait and avoid potential first mover disadvantage.
Often vendors have been the ones encouraging experiments. They want you to buy their latest product release and change your organisational processes to fit around their software. Analysts too are catalysts for experimentations, when they invent new terms and contribute to the hype. To be fair, some customers are also capable of dreaming up grand experiments on their own. Unfortunately for buyers, most internal managers can't tell an experiment from a tried, tested and proven solution. Vendors simply can’t say no!
I'm a big fan of learning, but there are many other ways to learn than experimenting. Next time you embark on a new online project, try removing all experiments and avoid the pain. Do proper planning and remember that experiments come with risks.
NB: Thanks to Gerry McGovern for inspiration with his 2002 article on Iterative design can be lazy design