The burden of proof... Sounds a bit too legal for project managers ? Let's go with a little story...
The story of a project manager who was confident and strongly believed that projects are about trust and teamwork, and who was getting stucked with his customer about a supply. This supply was a huge set of data.
Indeed, these data were a key delivery for the succes of the project, customer had full responsibility on it and, unfortunately, although target date was getting closer and closer, reporting on progress got fuzzier and fuzzier. Worried by this situation, our PM naturally gave a call to his counterpart to warn him about the envisioned risks and take actions to set the project back to a better course. Everything he got in return was an en bloc denial of the situation.
However, our PM experience was telling him that delivering such a massive amount of data which would have performed a full validation cycle (sources cleansing, extraction from sources, transformation, load, iterative dry-run and correct process until reaching an appropriate level of quality) could take up to several months. Delivering in far less time could potentially be achieved by poor quality deliverable. That was not acceptable either. Therefore, our PM wrote an e-mail as a reminder that scope of deliverable included a full validation of data sets prior to any delivery.
Now, our PM expectations were crystal clear and... Whatever ! Milestone was customer's responsibility after all ! In case of any delay, additional costs could be incurred to the party responsible for the delay. In case of poor data quality, milestone would be deemed as not achieved, and the additional costs would be inccured to the responsible too. That's it !
Still worried about this stucked situation, he talked about it with a senior executive who immediately answered:
"You cannot let the things run their course ! Even if this milestone is their responsibility, who is in charge of prooving the failure ?"
This bloodcurdling feedback striked our PM down. And senior executive went on:
"Milestone is under their responsibility, but YOU're responsible to ensure that the corresponding delivery is on time and of the expected level of quality. What if they do deliver a very poor quality data set on time ?
- Will you be able to audit data quickly to ensure they match your expectations on quality ?
- Will it even be feasible to audit data at all ?
- What about efforts and delay entailed by the audit ? It'll be charged on YOUR budget
- What about potential risks of discovering issues later on ? Once the milestone is accepted, YOU bear the risks"
That was enlightening !
So now, in any case, don't forget to wonder who's got the burden of proof then estimate and act accordingly !