Vendor Governance Placement

From WFM Labs


Vendor governance placement is the question of where, inside a client organisation, the functions of managing outsourced service delivery should sit. It is distinct from the question of what to outsource, and it is answered separately.

The short version of the method: commercial functions that require comparison across suppliers belong centrally; operational functions that require behaviour control belong with whoever owns the process; and commitment setting belongs to neither, because it requires information that no single function holds.

For the decision about which work is externally eligible in the first place, see Placement Rules and the Tenure Contract. For the design surface on which sourcing decisions sit, see Sourcing Design Axes: Node and Client Ownership.

A note on the evidence base, stated first

The specific placement question is not answered anywhere in the peer-reviewed literature. There is a mature body of work on governance mechanisms — contract design, trust, monitoring cost, control modes — and a separate, much thinner body on organisational structure. Almost nothing connects the two. No study compares a central vendor management function against embedded governance and measures an outcome.

What follows is therefore a derivation rather than a summary. Each placement decision below is traced to an evidenced principle, and the inference from principle to placement is the argument being made. Where the evidence runs out, that is stated rather than smoothed over.

One further limitation carries through everything. With few exceptions, the rigorous sources are drawn from information-technology outsourcing or from goods procurement. Applying them to labour-based service delivery is an argument by analogy that the literature itself has largely not tested — a point the framework's own principal authors have made in print (Lacity et al. 2011).

The principles that do the placement work

Six findings, each evidenced, together determine where a function belongs.

One. Monitoring cost is a function of proximity. Agency theory's information-asymmetry logic implies that whoever sits closest to a supplier's actual operational behaviour has structurally cheaper access to the information needed to monitor it (Eisenhardt 1989). Monitoring cost is not a matter of preference. It is set by distance from the work.

Two. Behaviour control requires process visibility, and is acquired rather than assumed. Control-mode research finds that behaviour control is added later and reactively, once process visibility has been obtained (Kirsch; Choudhury & Sabherwal 2003). The one study conducted in the business-process domain, across 234 paired projects, finds the relationship is symmetric: where client capability is weak, process control fails and outcome control works; where supplier capability is weak, the reverse (Liu, Wang & Huang 2017).

Three. Capability loss erodes the ability to manage at all. A study of 198 outsourcing relationships found that losing the underlying capability both degrades performance and independently erodes the client's ability to manage the relationship (Handley 2012). These are two separate effects. The second means a function that has lost the operational capability cannot manage the supplier regardless of the authority it holds.

Four. Standard-setting, measurement and enforcement are separable. Service-level agreement content decomposes into distinct foundation, change and governance dimensions (Goo, Kishore, Rao & Nam 2009; Goo, Kim & Cho 2010). They do not have to sit together, and treating them as one bundle is a design choice rather than a necessity.

Five. Gaming risk rises when the measured party controls measurement. The general finding that performance measurement produces dysfunctional behaviour is long established (Ridgway 1956), and the strongest evidence of a gap between reported and real performance comes from audited public-health data (Bevan & Hood 2006). The design consequence is that measurement requires independence from the party whose performance it reflects — which includes the client's own operational line, not only the supplier.

Six. Cross-supplier comparison requires a central view. Procurement organisation-design research finds that without a central programme, most organisations do not form an effective evaluation system at all, and that embedded governance fragments supplier intelligence and duplicates administration (Kanepejs & Kirikova 2018).

What this means in practice. Principles one to three push work outward toward the operational line. Principles four to six pull specific functions inward toward a central one. The split is not a compromise between them. It is what happens when each function is placed according to which principle governs it.

The three parties

Party What it uniquely holds
Commercial vendor function Cross-supplier visibility, contracting capability, and independence from delivery performance
Operational line Process ownership, and therefore the only available route to behaviour control
Planning function Demand information, and the error characteristics of its own forecasts

The third party is routinely omitted from vendor governance designs, which is the source of the commitment problem described below.

The placement

Function Accountable Also involved Governing principle
Sourcing strategy and supplier selection Commercial Operations, planning consulted Six
Contract terms and negotiation Commercial Operations on work specification Six
Measurement instrument definition Commercial Operations consulted Five, six
Threshold setting Commercial Operations consulted Four
Independent verification of reported performance Commercial — Five
Cross-supplier and cross-type scorecard Commercial — Six
Enforcement on breach Commercial Operations informed Four
Day-to-day quality application Operational — One, two, three
Performance and efficiency management Operational Commercial informed One, two, three
Job design and work specification Operational Commercial consulted Two, three
Ramp monitoring Operational Planning responsible for the schedule One, three
Supplier capability development Commercial Operations responsible for content Six
Demand forecast and error budget Planning — See below
Volume commitment setting Joint All three See below
Escalation Joint Posture over ownership See below

Why job design sits with operations, which is less obvious than it looks

Turnover in service work is driven substantially by how the job was specified rather than by who employs the person — roughly nine percent where roles are high-discretion and low-monitoring, against thirty-six percent where they are low-discretion and high-monitoring (Holman, Batt & Holtgrewe 2007).

A supplier asked to hold attrition down while being specified a low-discretion, high-monitoring job design has been given two incompatible instructions. Since the specification is written by whoever owns the process, the party that controls the strongest lever on turnover is the operational line, not the function that writes the attrition clause.

Why enforcement is separated from relationship management

Experimental work finds that penalty framing increases compliance effort while decreasing knowledge-sharing and commitment (Fehrenbacher & Wiener 2019). A contract's controlling function erodes trust while its coordinating function builds it (Lumineau 2017), and control provisions raise competence-trust while lowering goodwill-trust and reducing continued collaboration (Malhotra & Lumineau 2011, N=102 disputes).

Separating enforcement from the party that manages the relationship day to day is a way of containing that effect rather than eliminating it.

The counter-argument should be stated. The complementarity finding — that formal contracts and relational governance reinforce each other rather than substituting (Poppo & Zenger 2002; Cao & Lumineau 2015, meta-analysis of 149 studies and 33,051 relationships) — rests on contract customisation and relational investment being co-adapted by whoever manages the exchange. Separation may be what prevents that co-adaptation. The literature does not resolve this.

The commitment interface, where the evidence runs out

Volume commitments — committed volume or headcount for a period, with notice and minimums attached — are set commercially but are only meetable operationally, and are justified by a forecast produced by a third function that is often not in the room.

No literature addresses who should own them. The question sits between four bodies of work that have never been joined.

What is established

No contract type is universally preferred. Whether volume-based or capacity-based commitment is optimal depends on demand variability (Akşin, de Véricourt & Karaesmen 2008). Related work coordinates staffing level and service quality where supplier effort is hidden (Ren & Zhou 2008) and evaluates contract features under asymmetric information about supplier productivity (Hasija, Pinker & Shumsky 2008). This labour-native cluster is the correct anchor and is underused in practice.

Forecast error is larger than queueing models assume. Across four real operations, forecast error exceeds the natural stochastic fluctuation that staffing models treat as the dominant variance source, and models ignoring it systematically overestimate achieved service level (Steckley, Henderson & Mehrotra 2009).

Forecast skill has a short horizon. Standard methods beat a historical average only about two to three days ahead for intraday arrival patterns (Taylor 2008) — which is scheduling grain, not the weeks-to-quarters grain at which commitments are made.

What does not transfer

Supply-chain contract theory — quantity-flexibility contracts, options, capacity reservation (Cachon 2003; Tsay 1999) — is rigorous and built entirely on assumptions that do not hold for labour. There is no storability, so no salvage value, so buyback and revenue-sharing mechanisms have no direct analogue; severance is a sunk cost paid because labour cannot be stored, which is a different economic object. Retraining cost breaks fungibility. Legal notice and minimum-engagement terms create a cost floor rather than a foregone-salvage curve, which is the largest structural break. Ramp investment is sunk and not recovered on early termination.

Contract shape may transfer. Efficiency proofs do not, because they rest on the broken assumptions.

The named gap

Nothing in the literature ties forecast error by horizon to a defensible commitment horizon. This was searched for specifically and confirmed absent.

The practical consequence follows from what is established rather than from a study that tests it: a commitment set without the forecasting function present is being set against the wrong error budget. The party that knows how wrong the forecast is likely to be, at the horizon at which the commitment binds, is the planning function — and it is frequently not consulted.

Ramp-risk allocation is in the same position. Graduated relief across the first months of a ramp is the de facto convention, and no study correlates any allocation with measured ramp success. It is contractual convention, not studied fact.

How the placement changes across supply types

An estate running retained in-market delivery, owned remote centres and contracted supply simultaneously does not apply this design three times.

Column Contracted supply Owned remote centres Retained in-market
Contracting and enforcement Live Collapses — no counterparty Collapses
Measurement instrument and verification Live Live Live
Comparability scorecard Live Live Live
Operational performance management Live Live Live
Commitment setting Live Internal equivalent Internal equivalent

The row that must not collapse is measurement. Where the commercial column disappears because there is no contract, the temptation is to let the instrument and the verification disappear with it. If that happens, the estate can no longer compare a supplier against its own centres, which is the comparison every sourcing decision depends on.

This is the strongest argument for a central function that is not framed as vendor management at all. Its comparability role applies to internal delivery just as much, and framing it as supplier oversight is what causes the internal columns to be dropped.

Conditions under which the default reverses

The devolved default holds unless one of four conditions applies, each evidenced.

Supplier count beyond the visibility of a single operational unit. Beyond roughly five services per arrangement, cross-arrangement effects appear that no single unit can see (Bapna et al. 2023, N=49,057).

Weak enforcement environments, where trust and formal contract do not reinforce each other and the complementarity finding does not hold (moderators within Cao & Lumineau 2015).

Long-tenured relationships. No direct evidence exists for embedded managers becoming captured by the supplier they manage — the claim appears in practitioner material without support. But a well-evidenced analogue exists in auditor independence, where long tenure produces measurably greater leniency (Favere-Marchesi & Emby 2018, N=140, 76% against 39%). The relationship is structurally similar: long-tenured, information-asymmetric and judgement-dependent. The transfer is an argument, and should be presented as one rather than as a documented risk.

Where arm's-length enforcement has already visibly failed, which is a signal that outcome control alone is insufficient for that work.

Escalation, which behaves differently

Across thirteen paired conflict cases, the operative variable was posture rather than which function held the escalation — collaborative against adversarial — and several of the best outcomes were reached through collaborative commercial renegotiation (Lacity & Willcocks 2017).

This is the one place where the placement question is the wrong question. Escalation should be jointly owned and collaboratively conducted regardless of function, and adversarial or formal-dispute routing correlates with mutual dissatisfaction on both sides.

Failure modes

Placing performance management with a function that lacks process visibility. It can then exercise outcome control only, which is the weakest mode for discretionary work, and the capability to do better does not accumulate.

Letting the measurement column collapse for internal delivery. Removes the ability to compare supply types, which is the comparison that sourcing decisions rest on.

Allowing the measured party to control measurement. Applies to the client's operational line as much as to the supplier.

Setting commitments without the planning function present. Produces commitments made against the wrong error budget, and the error is discovered at the horizon where it is most expensive to correct.

Loading commitment governance onto penalty mechanisms. Increases compliance effort while reducing the information-sharing on which the next commitment depends.

Bundling standard-setting, measurement and enforcement because they are conventionally bundled. They are separable, and separating them is what allows enforcement to be centralised without centralising performance management.

Treating the design as fixed. Four named conditions reverse the default, and at least two of them — supplier count and relationship tenure — change over time without anyone re-examining the design.

Maturity Model considerations

At L1–L2, vendor governance is undifferentiated: one function holds contracting, performance and escalation, usually the one that signed the contract. Measurement is whatever the supplier reports.

At L3, contracting and performance management are separated, but measurement follows contracting rather than being independent, and internal delivery is not measured on the same instrument.

At L4, the placement above is broadly in force, measurement is independently verified, and the same instrument applies across supply types.

At L5, commitment setting includes the planning function as a matter of course, commitment horizons are set against a stated forecast error budget, and the four reversing conditions are reviewed rather than assumed stable.

Use this with Claude

A ready-to-deploy instruction set and reference files for building the placement engine are at Wiki:Packs/Placement Decision Engine Build (CP-WFM-008) — building the register, gate, cards and recompute log.

For analyzing vendor operating-model maturity results — including the routing-decision accountability test this article's placement principles act on — see Wiki:Packs/Vendor Maturity Assessment Analysis (CP-WFM-010).

See Also

References

  • Akşin, Z., de Véricourt, F. & Karaesmen, F. (2008). Call center outsourcing contract analysis and choice. Management Science, 54(2).
  • Bapna, R. et al. (2023). Multi-service outsourcing arrangements. N=49,057.
  • Bevan, G. & Hood, C. (2006). What's measured is what matters: targets and gaming in the English public health care system. Public Administration, 84(3).
  • Cachon, G. (2003). Supply chain coordination with contracts. Handbooks in Operations Research and Management Science, 11.
  • Cao, Z. & Lumineau, F. (2015). Revisiting the interplay between contractual and relational governance. Journal of Operations Management, 33–34. Meta-analysis, 149 studies, 33,051 relationships.
  • Choudhury, V. & Sabherwal, R. (2003). Portfolios of control in outsourced software development projects. Information Systems Research, 14(3).
  • Eisenhardt, K.M. (1989). Agency theory: an assessment and review. Academy of Management Review, 14(1).
  • Favere-Marchesi, M. & Emby, C. (2018). Auditor tenure and leniency. N=140.
  • Fehrenbacher, D. & Wiener, M. (2019). Effects of penalty framing in outsourcing contracts. Experimental, N=198.
  • Feeny, D.F. & Willcocks, L.P. (1998). Core IS capabilities for exploiting information technology. Sloan Management Review, 39(3).
  • Goo, J., Kishore, R., Rao, H.R. & Nam, K. (2009). The role of service level agreements in relational management. MIS Quarterly, 33(1).
  • Handley, S.M. (2012). The perilous effects of capability loss on outsourcing management and performance. Journal of Operations Management, 30(1–2). N=198.
  • Hasija, S., Pinker, E. & Shumsky, R. (2008). Call center outsourcing contracts under information asymmetry. Management Science, 54(4).
  • Holman, D., Batt, R. & Holtgrewe, U. (2007). The Global Call Center Report. ~2,500 centres, 17 countries.
  • Kanepejs, E. & Kirikova, M. (2018). Centralized vs. decentralized procurement: a literature review. CEUR Workshop Proceedings, 2218. Systematic review, 27 studies.
  • Lacity, M., Solomon, S., Yan, A. & Willcocks, L. (2011). Business process outsourcing studies: a critical review and research directions. Journal of Information Technology, 26(4).
  • Lacity, M. & Willcocks, L. (2017). Conflict resolution in business services outsourcing relationships. 13 paired cases.
  • Liu, S., Wang, L. & Huang, W. (2017). Effects of process and outcome controls on business process outsourcing performance. European Journal of Operational Research, 260(3). 234 paired projects.
  • Lumineau, F. (2017). How contracts influence trust and distrust. Journal of Management, 43(5).
  • Malhotra, D. & Lumineau, F. (2011). Trust and collaboration in the aftermath of conflict. Academy of Management Journal, 54(5). N=102.
  • Poppo, L. & Zenger, T. (2002). Do formal contracts and relational governance function as substitutes or complements? Strategic Management Journal, 23(8).
  • Ren, Z.J. & Zhou, Y.-P. (2008). Call center outsourcing: coordinating staffing level and service quality. Management Science, 54(2).
  • Ridgway, V.F. (1956). Dysfunctional consequences of performance measurements. Administrative Science Quarterly, 1(2).
  • Steckley, S.G., Henderson, S.G. & Mehrotra, V. (2009). Forecast errors in service systems. Probability in the Engineering and Informational Sciences, 23(2).
  • Taylor, J.W. (2008). A comparison of univariate time series methods for forecasting intraday arrivals. Management Science, 54(2).
  • Tsay, A. (1999). The quantity flexibility contract and supplier-customer incentives. Management Science, 45(10).
Retrieved from "https://wiki.wfmlabs.org/w/index.php?title=Vendor_Governance_Placement&oldid=3928"