“The best we can do is size up the chances, calculate the risks involved, estimate our ability to deal with them, and then make our plans with confidence.” – Henry Ford
- Estimation processes are about establishing confidence that the underpinnings of an endeavour have not only been professionally put together, but also do not expose an organisation to a level of risk it is not prepared to take.
- Estimation is an iterative process; expect estimates to be refined and change as knowledge, understanding and stability of scope, solution and delivery schedule increases.
- Estimating and pricing are different; never manipulate stable estimates of the effort to deliver a solution to a plan just to achieve the price a sales person wants – make scope changes instead.
- Create, mandate and use a standard set of principles and iterative estimation processes that have requirement Scope, Solution and Delivery Schedule at their heart because:
- Scope sets out what has to be produced and what the deliverables are to meet the requirement.
- Solution sets out the engineered solution for how the scope will be produced.
- Delivery Schedule sets out the plan of activities for how what needs to be produced will be done and by when.
- Estimates should always be produced by professional individuals with market sector and solution domain experience, using a clear and consistent methodology, and be scrutinised and open to challenge by others.
- Be clear from the outset what (e.g. software to be developed, volume of service calls) needs to be estimated, and what technique (e.g. lines of code, function points, published service catalogue) will be used, and how this will be translated into the cost to deliver.
- Use a published service catalogue founded on the total cost of your enterprise’s client-focused service operations for estimating standard service components.
- Never rely on one estimate or one estimator; always obtain at least one other completely independent estimate and understand and reconcile any differences.
- Always estimate effort based on average people, working normal office hours and allowing for sickness, vacations and national holidays; never estimate assuming delivery will be staffed by the best people.
- Compare or calibrate estimates against real metrics from similar types of work successfully completed by your enterprise where possible; however, apply intelligence in comparisons mindful that no two endeavours are truly the same.
- Treat estimates and assertions from subcontractors with scepticism and review these closely.
- Ensure estimates and delivery schedule are reviewed together to ensure the realism and sensibility of assumptions made; steep staff mobilisation or ramp down, and schedule parallelism or overlaps should always be checked for realistic real-world achievability.
- Always create a register of quantifiable risks and use this to guide contingency requirements; the impact of time delays not just labour, materials and expenses risks should be considered in setting an objective level of contingency.
- Where possible analyse estimates using an Industry parametric tool set (like SLIM) to assess the statistical confidence level against a database of Industry metrics; this can be useful for significant development or systems integration work.
- Remember – there is no such thing as perfect estimation! All estimates have an error bar regardless of the process, methodology or toolset used for their preparation.