82,883
edits
Playbook: Difference between revisions
no edit summary
Amwelladmin (talk | contribs) No edit summary Tags: Mobile edit Mobile web edit |
Amwelladmin (talk | contribs) No edit summary Tags: Mobile edit Mobile web edit |
||
| Line 1: | Line 1: | ||
{{A|negotiation| | {{A|negotiation| | ||
{{Image|Lear|jpg|A [[playbook]] yesterday.}} | |||
}} | }} | ||
A [[playbook]] is a comprehensive set of guidelines, policies, rules and fall-backs for the [[legal]] and [[credit]] terms of a {{t|contract}} that you can hand to the itinerant [[school-leaver from Bucharest]] to whom you have off-shored your [[master agreement]] {{t|negotiation}}s. She will need it because, being an itinerant school-leaver from Bucharest, she won’t have the first clue about the [[Subject matter expert|ISDA]] [[negotiation]]s, and will need to consult it to decide what do to should her counterparty object, as it certainly will, to any of the preposterous terms her [[risk controller|risk]] team has insisted go in the first draft of the {{t|contract}}. | A [[playbook]] is a comprehensive set of guidelines, policies, rules and fall-backs for the [[legal]] and [[credit]] terms of a {{t|contract}} that you can hand to the itinerant [[school-leaver from Bucharest]] to whom you have off-shored your [[master agreement]] {{t|negotiation}}s. She will need it because, being an itinerant school-leaver from Bucharest, she won’t have the first clue about the [[Subject matter expert|ISDA]] [[negotiation]]s, and will need to consult it to decide what do to should her counterparty object, as it certainly will, to any of the preposterous terms her [[risk controller|risk]] team has insisted go in the first draft of the {{t|contract}}. | ||
[[Playbook]]s derive from a couple of mistaken beliefs: One, that a valuable business can be “solved” and run as an [[algorithm]], not a [[heuristic]];<ref>This is a bad idea. See {{author|Roger Martin}}’s {{br|The Design of Business: Why Design Thinking is the Next Competitive Advantage}}.</ref> and two, that, having been solved, it is a sensible allocation of resources to have a cheap, uninformed human being run that process rather than a machine.<ref>Assumption two in fact falsifies assumption one. If it really is mechanistic, there is no reason to have a costly, capricious human “helping to manage” — i.e., ''interfering in'' the process.</ref> | [[Playbook]]s derive from a couple of mistaken beliefs: One, that a valuable business can be “solved” and run as an [[algorithm]], not a [[heuristic]];<ref>This is a bad idea. See {{author|Roger Martin}}’s {{br|The Design of Business: Why Design Thinking is the Next Competitive Advantage}}.</ref> and two, that, having been solved, it is a sensible allocation of resources to have a cheap, uninformed human being run that process rather than a machine — which is in turn also a bad idea, just a less bad one than using a human.<ref>Assumption two in fact falsifies assumption one. If it really is mechanistic, there is no reason to have a costly, capricious human “helping to manage” — i.e., ''interfering in'' the process.</ref> | ||
In {{author|Thomas Kuhn}}’s argot<ref>{{br|The Structure of Scientific Revolutions}}. Brilliant book. Read it. </ref> playbooks | In {{author|Thomas Kuhn}}’s argot<ref>{{br|The Structure of Scientific Revolutions}}. Brilliant book. Read it. </ref> playbooks deal with situations of “[[normal science]]”: They map out the discovered world. They contain no mysteries or conundrums. They represent tilled, tended, bounded, fenced, arable land. Boundaries have been set, tolerances limited, parameters fixed, risks codified and processes fully understood. | ||
[[Playbook]]s are [[algorithm]]s for the [[meatware]]: they maximise efficiency when operating within a fully understood environment. They are inhabited exclusively by [[known known]]s. No [[playbook]] will ever say, “if the counterparty will not agree this, make a judgment about what you think is best.” All will say, “any deviations from this requirement must be approved by [[Litigation]] and at least one [[Credit]] officer of at least C3 rank.” | [[Playbook]]s are [[algorithm]]s for the [[meatware]]: they maximise efficiency when operating within a fully understood environment. They are inhabited exclusively by [[known known]]s. No [[playbook]] will ever say, “if the counterparty will not agree this, make a judgment about what you think is best.” All will say, “any deviations from this requirement must be approved by [[Litigation]] and at least one [[Credit]] officer of at least C3 rank.” | ||
| Line 12: | Line 12: | ||
As far as they go, [[playbook]]s speak to the belief that, as [[normal science]], ''the only material [[risk]] lies in not complying with established rules'': They are of a piece with the [[doctrine of precedent]]: when they run out of road, one must appeal to the help of a higher authority, by means of [[escalation]] to a [[control function]], the idea being (in theory, if not in practice) that the [[control function]] will further develop the [[algorithm]] to deal with the new situation, the same way the courts of the [[common law]] do — ''[[stare decisis]]'' — and it will become part of the corpus and be fed back down into the playbook of established [[process]]es.<ref>This rarely happens in practice. [[Control function]]s make ''[[ad hoc]]'' exceptions to the process, do not build them into the playbook as standard rules, meaning that the [[playbook]] has a natural sogginess (and therefore inefficiency).</ref> The [[algorithm]] operates entirely ''inside'' the organisation’s real {{tag|risk}} tolerance boundaries. This is a good thing from a risk monitoring perspective, and is inevitable as a matter of organisational psychology — [[if in doubt, stick it in]], as [[Casanova’s advice|Casanova]] used to say — but it all comes at the cost of efficiency. The [[escalation]]s it guarantees are a profoundly [[waste]]ful use of scarce resources. | As far as they go, [[playbook]]s speak to the belief that, as [[normal science]], ''the only material [[risk]] lies in not complying with established rules'': They are of a piece with the [[doctrine of precedent]]: when they run out of road, one must appeal to the help of a higher authority, by means of [[escalation]] to a [[control function]], the idea being (in theory, if not in practice) that the [[control function]] will further develop the [[algorithm]] to deal with the new situation, the same way the courts of the [[common law]] do — ''[[stare decisis]]'' — and it will become part of the corpus and be fed back down into the playbook of established [[process]]es.<ref>This rarely happens in practice. [[Control function]]s make ''[[ad hoc]]'' exceptions to the process, do not build them into the playbook as standard rules, meaning that the [[playbook]] has a natural sogginess (and therefore inefficiency).</ref> The [[algorithm]] operates entirely ''inside'' the organisation’s real {{tag|risk}} tolerance boundaries. This is a good thing from a risk monitoring perspective, and is inevitable as a matter of organisational psychology — [[if in doubt, stick it in]], as [[Casanova’s advice|Casanova]] used to say — but it all comes at the cost of efficiency. The [[escalation]]s it guarantees are a profoundly [[waste]]ful use of scarce resources. | ||
In theory the [[control function]] will have its own playbook, and the “court of first instance” is as bound by that as the baseline process is by the basic playbook. There is an [[algorithm]], a recipe, and the main ill that comes about is by not following it. Hence the existence of an [[internal audit]] function. | In theory the [[control function]] will in turn have its own playbook, and the “court of first instance” is as bound by that as the baseline process is by the basic playbook. There is an [[algorithm]], a recipe, and the main ill that comes about is by not following it. Hence the existence of an [[internal audit]] function. | ||
And are we even going to talk about the fact that the big shock risks that hit the systems are never ones that have previously been recognised, analysed and subjected to constant monitoring? [[Black swan]]s gonna be [[black swan]]s, yo. | And are we even going to talk about the fact that the big shock risks that hit the systems are never ones that have previously been recognised, analysed and subjected to constant monitoring? [[Black swan]]s gonna be [[black swan]]s, yo. | ||
===Playbooks, design and user experience=== | |||
Let us take it as a given that the PlayBook methodology of using set instructions to delegate administrative tasks to unskilled personnel, is fit for purpose only within the bounds of normal operating conditions. | |||
For example, in a negotiation playbook, risk control department A has stipulated a starting position X, but has recognised that if a client does not agree to X, a satisfactory compromise may be found at Y. The playbook accordingly “empowers” the negotiator to offer that compromise. Only if client should also not agree to Y will there be an [[escalation]], back to a risk control department A whom, no doubt following its own interior playbook, may sanction a further derogation from standard terms at Z. There may then follow an extended firefight between senior risk personnel on either organisation — albeit conducted through there are uncomprehending negotiation personnel in Bratislava — which will culminate at final agreement at position Z'. | |||
By codifying this process, so the argument goes, not only may we engage materially cheaper negotiation personnel, but we effectively triage our client base and at the same time improve our systems and controls over the previous process whereby a more experienced negotiator made it up as she went along. | |||
We have certainly added to our systems and controls; no doubt about that. | |||
But look at this from about: risk control department is stipulating position X where in reality it will accept position Z or even Z'. The playbook, and all those wonderful systems and controls, in play only for the the portion of the negotiation between X and Y, being a position a mile behind enemy lines. | |||
No doubt middle management will regale any steerco or opco that will listen about the magnificent management information and statistics it has about the negotiation process. But all these gears are involved, and all all the systems is running over a part of the process that presents zero risk to the organisation. Negotiation for ''all'' clients takes longer, and now the portfolio will be distributed over a range of points between X and Z, making practical control portfolio more difficult. | |||
The client will not enjoy the negotiation process any more than you will, if you are presenting your client with its first experience of your organisation as the image of impersonal bureaucracy. With the greatest respect for our friends in the Slovak Republic, your new clients would prefer to be hand-held by a salesperson and in London than a school leaver in a call centre in Bratislava.<ref>Granted, in this day and age, your client is almost certainly housing its negotiation probability out of Acorn centre in Bratislava too. </Ref> | |||
What to do? Negotiation difficulties will generally fall into two camps: 1, the client (or its negotiation team, which also has been outsourced to Bratislava) won't ''understand'' your document and will therefore insist on changing it. 2, it will find your legal terms fundamentally unreasonable. | |||
Both scenarios are likely; often at once: it’s highly likely people in your own organisation won't understand your documents, so it is a bit rich expecting your clients to.<ref>Best example is the [[hypothetical broker dealer]] valuation terms in a [[synthetic equity swap]].<> The answer to both lies in product [[design]] and consideration of [[user experience]]. | |||
For the confusing and misunderstood x, the answer is straightforward, but difficult: ''simplify'' them. This is a matter of not just language, but logical structure, though by simplifying language often convoluted logical structures become easier to see and therefore easier to fix. And while capital markets drafting is famously dreadful, there is emerging technologies that may help. | |||
Run your templates through a GPT-3 engine and ask it to simplify them. It almost certainly won’t be perfect and almost certainly will make errors, but ''it is free''. Checking for errors and running quality control is what ''you'' are there for. It will break the back of an otherwise impossible job. | |||
{{sa}} | {{sa}} | ||