Shared Services for iGaming Groups

The service catalogue, by relationship type

Service	Relationship	The contract's core clause
Compliance / RG / risk	Mandatory authority with service-grade operations	Response SLAs by severity; the brand obeys outcomes, per the authority structure
Payments & treasury	Mandatory infrastructure	Provider onboarding lead times published; approval-rate reads per brand-market cell
Data platform & BI	Mandatory core, self-service surface	Freshness SLOs per mart; brands build on the dictionary, never beside it
Platform engineering	Shared roadmap	The published queue and prioritisation rule — see brand vs platform
CRM operations	Service with brand steering	Brands own strategy and voice inside the gated rule engine; the service owns execution capacity
Creative & content ops	Optional service	The one tier where a brand may genuinely go external — and the hatch keeps the internal shop honest

The contract's anatomy

Catalogue: what is offered, what is explicitly not, and the request channel — one intake, not five DM threads.
Response classes: severity tiers with turnarounds, so the compliance deadline and the cosmetic tweak never share a queue position.
Published performance: the service's own SLA attainment, visible to its customers — the same freshness-on-the-surface honesty as the dashboards.
The allocation key: stable, usage-weighted where measurable, boring by design.
The escalation path: documented, to the portfolio owner, with SLA evidence — the hatch that keeps monopoly quality at bay.

Self-service as the scaling strategy

The mature shared service grows by removing itself from loops rather than by adding coordinators: brand-configurable parameters inside guardrails (the configuration matrix pattern generalised), templated request types that execute without human routing, and dashboards that answer the questions tickets used to ask. The heuristic from the FAQ bears repeating as the operating rule: any request type recurring monthly is a self-service feature with a backlog entry.

The two decay modes to watch

The monopoly drift: captive customers, no published SLAs, queue opacity — detected by rising informal workarounds (brands hiring shadow ops, the org-design page's bypass diagnostic). The vendor cosplay: internal invoicing ceremony, negotiation overhead and chargeback disputes consuming more energy than the service delivers — detected when brand leads discuss allocation more than customers. Both decay modes have the same antidote: measure the service like a product (adoption, satisfaction, SLA attainment) and review it in the portfolio rhythm like any other investment.

Continue reading: Brand vs platform ownership — the deepest seam, refereed. The org design — the frame these services sit in.

Shared services — common questions

What turns a central team into a shared service?

A service contract: what the function provides, at what turnaround, measured how, requested through which channel — published to the brands like a vendor's terms. The distinction is behavioural, not semantic: a central team does work it chooses in the order it chooses; a shared service delivers a defined catalogue against defined SLAs, and its performance is as visible to its internal customers as any external vendor's would be.

Should brands pay for shared services internally?

Usage-based allocation, yes; internal invoicing theatre, rarely worth it. The allocation exists to keep brand P&Ls honest and consumption rational (a free service gets consumed like one), not to simulate a market. The rules that matter: the key is stable, published and boring; heavy users pay more; and the allocation never becomes a negotiation surface that distracts brand leads from running brands.

Which functions must never be 'services' brands can decline?

Compliance, protection and risk. A brand cannot opt out of the RG function or shop for lighter-touch fraud rules — those are the group's licence surface, and their authority is mandatory, per the org-design page. The service framing still helps operationally (queues, SLAs, published processes), but the relationship is regulator-proxy, not vendor-customer: the brand consumes the service and obeys the function.

What is the escape hatch, and why does it keep quality?

A defined path for a brand to escalate persistent service failure — to the portfolio owner, with evidence against the published SLAs — and, for genuinely commodity services, the theoretical option of an external alternative. The hatch's value is almost never its use; it is the discipline its existence imposes. Shared services with captive customers and no escalation path drift toward the quality level of any monopoly.

How do shared services avoid becoming the bottleneck?

By industrialising with brand count: self-service before headcount (the brand configures within guardrails instead of filing a ticket), published capacity and a visible queue (the roadmap discipline applied to operations), and tiered response classes so a compliance deadline never queues behind a cosmetic request. The scaling page tracks where each of these becomes necessary; the short version is that every ticket type that recurs monthly is a self-service feature waiting to be built.

Shared services, run as products.

The service catalogue, by relationship type

The contract's anatomy

Self-service as the scaling strategy

The two decay modes to watch

Shared services — common questions

See it on a live stack.