People, process, technology. In that order.

CRM is a business philosophy before it is a piece of software. I've watched brilliant platforms produce disappointing projects, and the autopsy always finds the same cause. The order was wrong.

The wrong order

The typical sequence goes like this. The licence is bought. The integrator is chosen. The go-live date is announced. Then, somewhere around the second steering committee, someone asks what problem the system is supposed to solve.

I once stopped an entire RFP because the client couldn't answer that question. Awkward meeting, excellent outcome. Months of implementation would have automated a question mark.

The right order

People first. Who owns the customer relationship today? Who loses influence when data becomes shared? A CRM redistributes power, and people defend their positions with impressive creativity. Skip this conversation and adoption dies quietly in the field, one unfilled record at a time.

Organisation next. Territories, targets, responsibilities. If two departments fight over one customer, the system inherits the fight and gives it better reporting.

Then process. Map what actually happens, and repair what's broken. Automating a broken process gives you the same process, faster (which is worse).

Technology last. By then it really is the easy part. Choosing clouds, objects and integration patterns is craft, and I love the craft. But craft can't rescue a project that skipped the first three steps.

"Reverse the order, and no licence will save the project."

What I do about it

Every engagement starts with the same questions, in the same sequence. Who will use this? How is the organisation arranged around the customer? What does the process look like on a Tuesday afternoon, as opposed to in the slide deck? Only then: which technology, if any.

So when someone asks me which cloud to buy, my first question is who's going to use it, and my second is what for. Technology last. That order has never failed me. The reverse fails on schedule.