We are rather good at introductions and less good at the paragraph that says what will be delivered on which Tuesday. A work package is that paragraph, with a budget band and a person who will answer the telephone.
Blighty Boffins publishes work packages because search is necessary and not sufficient. The expert is found. Then someone must say what the work is.
If your scope contains the word 'explore' more than twice, go for a walk and try again.
If your scope contains the word 'explore' more than twice, go for a walk and try again.
The test we apply is simple and slightly brutal. Could a competent stranger read your package and know whether they are the right person for it? Could they price it? Could they tell, without asking, what would count as having done the job well? If any of those answers is no, the package is not ready, and publishing it will produce applications from people who are guessing.
The commonest failure is a scope that describes a subject rather than a task. 'Investigate the potential of advanced materials for lightweighting' is a subject. 'Establish whether a magnesium alloy bracket meets our stiffness requirement at a mass below 340 grams, and say what it would cost to make fifty' is a task. The second attracts three serious applications. The first attracts eleven vague ones and a consultancy.
The second failure is silence about money. A budget band is not a commercial weakness; it is the single most useful filter on the page. Applicants who cannot do the work within it will not apply, which saves everybody a fortnight of courteous correspondence.
The third, and the one we find most instructive, is the absent constraint. Every real piece of work has an awkward fact: the site can only be visited on a night shift, the samples are irreplaceable, the chamber is booked until March, the data is handwritten for four months of the record. Publishers instinctively hide these, believing they will deter applicants. They do deter some — the ones who would have discovered the constraint in week six and asked for a variation.
None of this requires much writing. The best packages on our board are shorter than the average non-disclosure agreement and considerably more useful. They were simply written by someone who had decided, before they started typing, what they actually wanted.