I'll start off with a complete agreement with Cyntihia's first sentence. "Earned Value has its limitations." And to go further, the cost-centric view of EV and the questionable promises of the approach can lead to the very kind of quandary that she finds herself in.
I don't want to get off on a rant here, but . . .
I've been sitting on this thread for a few weeks now, trying to come up with a civil response. It's difficult enough to get beyond the idea of discussing "accurate estimates" or forecasts on a supposedly professional PM conversation without gagging. But to listen to an expectation of 100% accurate forecasts of project performance even only 30 days out is downright nauseating. (I won't even go into the question of using past (even EV-based) performance to predict future events.)
A friend of mine recently said "Tell me how you'll measure me, and I'll tell you what damned stupid things I'll do to make that measurement look good."
If your job is (and if you are going to be judged on) predicting costs with 100% accuracy, I'm sure you can find good creative ways to do just that. But I would predict that the time, effort, and attention to the manipulation of costs and cost reporting will severely impact your ability to address what is really important in the project management context.
One way of doing so would be to put plenty of safety in your estimates, then use that safety whether it's needed or not. That's about the easiest way of doing what is wanted, but I'm sure everyone reading this thread can predict what that will do to your project lead times and to the ability to get more real work done.
This is one of those situations that force me to ask what the goal of your PM effort is. Is it to meet financial estimates, or is it to get your e-commerce sites up and running as quickly as possible? What contributes more value to the organization, controlling costs to the penny or turning on revenue-generating or process-improving systems? The inane request for 100% accuracy at a project level demonstrates an all-too common "penny/cost wise - pound/benefit foolish" attitude. (I'll bet that concurrent with this attitude is a perceived need to keep everyone busy, which leads to cross-project interference and multi-tasking, which leads to things taking longer than planned, which leads to added unpredictability. I'll bet there are also pressures to launch projects before the resources are prepared to take them on as well, further exacerbating the situation.)
I have to admit that looking at Cynthia's profile, and with an awareness of the financial straits that the company mentioned in her email address is in these days, I can understand where this penny-pinching attitude might be coming from. But I think the questions in the previous paragraph are still valid. If the portfolio of projects is appropriately chosen for real value, the faster they get implemented, the sooner the company will get real benefits that should far outweigh costs to implement. If the projects are not truly valuable to the bottom line, then I can understand the cost focus. If they are, "damn the cost, full speed ahead!"
Come to think of it, I would expect that in aggregate your total project costs are pretty well fixed (or can be) for the next month out, so at a portfolio-wide level, the demand for accuracy is really not a problem. The only problem is that unless someone repeals Murphy's Law, individual projects will vary in the ability to predict what gets done in that period.
If you're forced, promise costs at an aggregate portfolio level, and forget them. Then track the projects with an eye on what matters -- the schedule -- and use the resources where they best contribute to moving the projects to completion and creating value for the company. Cost is merely an effect of schedule and scope once a schedule and scope have be agreed to. You can't maange effects -- you can only manage the causes. Assuming scope is fixed, that leaves you with the schedule. Manage it and costs will be what they need to be. And whatever you do, don't rely on some bogus formula based on past performance to predict future events. Ask the resources regularly, frequently, and directly what their time estimate to complete their task is. Don't accept answers like "we're on track." Demand numbers that reflect reality. Then and only then will you have information useful for reasonable prediction.
Your job is not, can not, and should not be to predict anything perfectly. It is to provide information and guidance regarding options and strategies for dealing with the variation that will exist, no matter what the estimate or forecast says.
. . . Of course, that's just my opinion. I could be wrong.
