Estimating effort and cost in an IAM bid

Identity programmes touch a lot of surfaces: source systems, directories, joiner mover leaver processes, application onboarding, privileged accounts, and the approvals around all of it. Estimates drift when that surface is summarised as a single number instead of being broken down.

The other common cause is unrecorded assumptions. If the estimate assumed three source systems and the client has nine, that difference only becomes visible in delivery unless it was written down at bid time.

Estimate from the things that actually change the effort: how many applications are in scope, how many identity sources feed the platform, how many roles or entitlement models need designing, the state of existing data quality, and the number of environments and approvals involved.

Naming those drivers makes the estimate reviewable. A reader can disagree with a driver value without rejecting the whole number, which is a far more productive conversation.

Convert effort to cost with a rate card by role rather than a blended average. An engagement that leans on architects and privileged access specialists costs differently from one that is mostly configuration work, and a blended rate hides that difference from your own margin review.

Keep the card in one place and reuse it across bids, so two teams pricing similar work reach comparable numbers.

Group the effort into work packages with deliverables, roles, dependencies, and risks. This does three things at once: it gives the client a delivery narrative, it lets you check that nothing is missing, and it makes phasing negotiable if the budget comes back short.

Dependencies matter as much as totals. A package that cannot start until the client completes a data cleanup is a schedule risk worth stating in the proposal.

BidGenX derives a day and cost estimate from the complexity drivers it finds in the tender documents and your own per-role rates, then builds a costed, dependency aware delivery plan with deliverables and risks per package. Every figure stays editable, and the assumptions behind it stay visible.