The last big agency I worked at had a person work full-time for months to do a statistical analysis of our historical work, to determine what constituted a Small, Medium, and Large project. These were ranges (and they were for the pharmaceutical industry, so scale anything you believe a project should cost by 1,000). The analysis looked over projects that had gone well, and projects that hadn’t, and came up with some “tells” of what to look out for, and a scaling factor to apply when they were seen. This made estimating a project much less of a fire drill, and allowed for a better outcome. It was a serious investment, though, and relied on the company having been in business, doing the same thing, since the Web became a commercial playground.
Here’s how this got applied in practice. When a customer would send an RFP or request a bid, we would have a top-level conversation with the client about their goals and their vision of the project. We would write a proposal outlining the project as we saw it, noting any potential speed-bumps or areas of fuzziness, and then tell them which bucket we thought their project was in. This cleared out a lot of projects where we were just one of “the three bids”, rather than the vendor the client actually wanted. It also scared off the smaller jobs, because a “small” job for MBC would pay my mortgage for a year.
If the client didn’t spit out the hook at this point, we would go back over the outline, try to refine the unknowns, and come up with a fully-qualified estimate, accurate within 15%. But here’s the key. If the client wanted that, they would have to pay us to create it, up front, and that cost was automatically 20% of the total project “bucket” cost. So instead of doing a lot of spec work to get the job, we would get the client to pay for the project plan up front. The sweetener we offered them was that at the end of that phase, if they wanted to walk, they could have the entire plan and all our thinking that went into it. As far as I know, nobody ever walked.
On Jan 20, 2014, at 2:12 AM, RavenManiac wrote:
I guess there are several ways to approach this, and to be honest I haven’t found the perfect approach yet.
offtopic mailing list
Update your subscriptions at: