Designing an Institutional ODD Operating Model
Most conversations about ODD operating models start with capacity. How large should the internal team be? What should be outsourced? How often should managers be reviewed? What can technology take over?
These are necessary questions, but they are design choices rather than the design principle itself.
Operational DD exists to help an investor form reliable judgement about an organisation to which it delegates capital, despite never having complete visibility into how that organisation operates. The ODD output then has to be usable within the investor's own governance process. An effective operating model should be built backwards from where this output is being used in decision-making.
Two institutions can therefore organise ODD very differently and both have credible models. The important issue is whether routine work, specialist analysis, experienced judgement and institutional accountability sit in the right places.
A good operating model therefore does not simply maximise reviews per person. It keeps routine work from consuming the people needed when interpretation becomes difficult.
Protect scarce judgement capacity
In practice, signs that an ODD function is overstretched often emerge before any deadlines are missed
Reviews are still being completed, but the same senior professionals increasingly become the final stop for routine questions, marginal issues and quality checks. Their calendars fill up long before the review statistics show an obvious capacity problem.
That matters because experienced judgement capacity is usually the least scalable part of ODD. Documents can be processed more efficiently, information can be structured and technology can remove administrative work. And even external specialists can increase delivery capacity. None of those things removes the difficult moments when the evidence does not fit neatly together.
An asset manager’s organisation chart may place responsibility clearly with one function while interviews suggest that effective authority sits elsewhere. A control may appear well designed while evidence of operation remains thin. Several individually modest observations may begin to point towards a wider dependency. These are the situations where senior attention adds the most value. Lestrade's own framework treats difficult interviews, conflicting evidence and structural dependencies as the work for which experienced judgement capacity should be protected.
A good operating model therefore does not simply maximise reviews per person. It keeps routine work from consuming the people needed when interpretation becomes difficult.
Keep institutional ownership where it belongs
External specialists can perform a substantial part of the investigative and analytical work. They can review evidence, conduct interviews, challenge management and form an independent professional opinion.
Some responsibilities should nevertheless remain clearly with the investor.
Risk appetite is institutional. So are escalation authority and the final decision to delegate capital. Internal stakeholders also need enough understanding of material ODD conclusions to challenge them and relate them to the wider investment case. This retained boundary is explicit in the underlying operating-model thinking.
The distinction becomes clearest when difficult conclusions emerge.
An external reviewer may determine that a governance dependency is material and only partly mitigated. What happens next can reasonably differ between investors. The significance of the capital allocation, investment conviction and the institution's own tolerance for the residual exposure all matter.
The provider can sharpen that decision. It should not quietly become the owner of it.
A useful warning sign is when nobody inside the institution can explain an important ODD conclusion without asking the provider to join the conversation. At that point, outsourced work may have started to become outsourced understanding.
Good ODD outsourcing should leave the institutional investor with more capacity for its own judgement, not less ability to exercise it.
External work should be challenged, not recreated
A different problem arises when institutions outsource the work but not the workload.
The provider completes the review, after which the internal team re-reads much of the documentation, repeats key questions and reconstructs the analysis before it is willing to rely on the conclusion.
Some internal challenge is essential. Re-performing the diligence defeats much of the point.
The stronger model is one in which the provider's work is transparent enough to interrogate without recreating it. Internal stakeholders should be able to understand why a conclusion was reached, what evidence supports it and where uncertainty remains.
This is a demanding standard for the external provider. A polished conclusion with no clear route back to the evidence may be persuasive, but it creates dependence on the specialist's authority rather than confidence in the judgement itself.
Good ODD outsourcing should leave the institutional investor with more capacity for its own judgement, not less ability to exercise it.
Build an ODD function that remembers
Point-in-time reviews are only part of the job.
A remediation delay may be entirely understandable in isolation. The third delay around related issues can mean something different. A senior departure may be manageable until it becomes part of several years of turnover. A governance weakness accepted during the previous review may deserve more attention after the organisation has doubled in scale.
History changes interpretation.
A mature ODD function should therefore remember not only what it concluded, but also why it came to this conclusion. Which risks were consciously accepted? What evidence supported the judgement? What remediation was expected? Which developments were supposed to be monitored?
That institutional memory should not live mainly in the heads of a few experienced reviewers, whether internal or external. The history of the delegation belongs to the investor and should remain connected to its own governance record.
Monitoring and subsequent reviews then become continuations of an existing judgement rather than repeated attempts to rediscover the manager from the beginning.
This is an important dividing line between conducting a series of ODD reviews and operating an ODD function.
Design backwards from the decision
The final test is practical.
Give the completed ODD output to a senior decision-maker who did not participate in the review. They should be able to understand what matters, why it matters, how securely the conclusion is supported and whether anything needs to happen as a result.
If the ODD team routinely needs another twenty-minute explanation before its report becomes usable, the analysis may be excellent, but part of the operating model has not completed its job. Institutional usability is itself a measure of ODD quality.
The same principle applies to the mechanics around the review. Managers need coordinating, information needs to move efficiently, and conclusions need a route into the investor's governance process. Rigour should not require unnecessary operational burden; Lestrade's positioning deliberately treats practical integration and low-friction execution as part of ODD quality rather than as service features around it.
There is no universal formula for the balance between an internal team, external specialists and technology. Nor should there be.
A well-designed ODD operating model makes reliable judgement easier to produce, easier to challenge, easier to retain and easier to use. Headcount, outsourcing and technology should serve that outcome rather than become substitutes for it.
Reviewing your ODD operating model? Talk to Lestrade.
