The way I see it, the past has been a gold mine for the likes of the Big 5, and their implementations at Fortune 50 companies. These projects typically have large consulting staffs and large reengineering efforts involved.
While there is still demand within these companies, the much larger market is within small and mid-size companies, which cannot sustain the type of project the Big 5 like to deliver.
These projects will require a much different approach and more activities assigned to a smaller team of consultants. These consultants must have a stronger, more diverse skill set to be effective on these projects.
Also, small and mid-size companies will be utilizing pre-configured solutions and ASP's to deliver them, often cutting out the need for Big 5 type implementations.
I also think the "boutique" consulting shops catering to mid-market and small-market clients will be the real survivors. I would also add that those consulting shops willing to work under "Fixed-cost" and "Phased billing" contracts will also thrive. \
I am not familiar with SAP products, but I am an Oracle IT/end-user for a small Oracle shop. We only work with Oracle consulting firms willing to give us "fixed-bid" and "phased billing". Using Big 5...forget it.
Also, are there small shops utilizing SAP? I am aware of some large manufacturing companies locally that use SAP. However, I'm not aware of any small to mid size companies. At any rate, the perception here is that only large companies can afford to purchase and support SAP. I'm not sure if this is an invalid perception, but I do know that locally, there are many companies using Oracle Applications...from small shops like mine to large Fortune 100 companies.
I agree with Across04's assessment of future ERP projects. The next generation ERP will be "point-solutions". Those ERP vendors who can de-couple their products and sell "by the piece" will be the survivors. All companies are getting ERP smart. Instead of spending 100's of millions in full-scale ERP, they want to see return on 5-10 million $ point-solutions. Just my opinion.