TL;DR
For a holding company deploying CRM across multiple trading entities, implementation costs typically run at 1x to 3x first-year licence fees, with McKinsey data showing large software projects overrun budgets by 66% on average. A five-company group paying £48,000 annually in licences should budget £144,000 to £208,800 for implementation, integration, permissions logic and group reporting, then ongoing change requests at £800 to £1,500 per day. Licences are 40% to 75% of total three-year cost depending on complexity. Post-launch, internal admin time (60% of a sales rep's day per Salesforce research), custom reports commissioned per board question, and per-entity additions drive recurring costs that dwarf the seat price.
Every CRM proposal a finance director reads has one number printed with total confidence and every other number left as an estimate, a rate card or a footnote. The confident number is licences. CRM total cost of ownership, the frame an FD actually thinks in, is dominated by the lines underneath it: configuration, integration, data migration, and the permanent drip of change requests that starts the week after go-live. For a group of companies, that gap between the quoted number and the owned number is wider than it is for a single trading entity, and the reason is structural rather than anyone's fault.
Executive Summary
CRM total cost of ownership is the sum of licences, implementation services, integration and migration work, internal staff time, and post-launch change requests over the life of the system. For a holding company deploying across multiple trading entities, implementation costs typically run at 1x to 3x first-year licence fees, so a group paying £48,000 annually in licences should budget £48,000 to £144,000 for the initial deployment alone. Licences are the only line with a published price, which is why they anchor the conversation and why they mislead. Gartner's own guidance on estimating a CRM project tells application leaders to work out which multiple of the licence fees they are likely to spend on consulting, which tells you where the weight sits. McKinsey and the University of Oxford, studying more than 5,400 IT projects, found large IT projects run 45% over budget while delivering 56% less value than predicted, with large software projects worse still at 66% over budget. Multi-entity groups pay a premium on every one of those lines because the platform underneath them assumes one company, and every departure from that assumption is billed as bespoke work.
What actually sits inside CRM total cost of ownership?
A CRM contract has roughly nine cost lines. Two of them are quoted precisely, three are quoted as ranges, and four are almost never quoted at all, which is where the budget variance lives.
| Cost line | Quoted up front? | Who really pays it |
|---|---|---|
| Per-user licences | Yes, to the penny | IT or group finance budget |
| Implementation and configuration | Yes, as an estimate against a rate card | Project budget |
| Data migration and de-duplication | Partly | Project budget, then overrun |
| Integrations to finance, ERP and marketing | Sometimes, per integration | Project budget, per entity |
| Change requests after go-live | No | Operating budget, indefinitely |
| Group and board reporting | No | The FD's own time, plus partner day rates |
| Internal time from sales ops and admins | No | The sales P&L, invisibly |
| Renewal uplift and new-entity additions | No | Procurement, at renewal |
| Exit, extraction and re-migration | No | Whoever inherits the decision |
The four unquoted lines share a characteristic. None of them can be sized during a sales cycle, because none of them depend on the product. They depend on the distance between how the platform assumes a business works and how this particular group actually works. That distance is discovered during implementation, priced in hours, and re-priced every time the group changes shape.
Licences are the price of the software. Everything else is the price of the gap between the software's assumptions and your operating model.
What's the typical implementation cost when deploying a CRM across multiple trading entities within a holding company?
Implementation costs for a multi-entity group typically range from 1x to 3x first-year licence fees, covering configuration, permissions, data migration, integrations and group reporting. McKinsey's study of more than 5,400 IT projects found large software projects overrun budgets by 66% on average, so a realistic budget for a five-company group paying £48,000 in annual licences is £96,000 to £208,800 for the initial deployment. That figure excludes post-launch change requests, which accrue at consultant day rates of £800 to £1,500 indefinitely.
Why is the implementation bill usually larger than the licence bill?
Mainstream CRM platforms are built generic to address the widest possible market, which means the fitting happens later, in configuration, at a services day rate. For a single trading company, that configuration work is bounded by the platform's standard data model. For a multi-entity group, it becomes structural: cross-entity permissions, consolidated reporting and the ability to share accounts between brands while maintaining separation elsewhere require custom logic that sits outside the product roadmap and is priced as bespoke work. That is the mechanism behind Gartner framing early CRM cost estimation as a multiple of licence fees rather than as an addition to them.
The overrun risk on that services line is well documented. McKinsey's work with Oxford put the average cost overrun on large IT projects at 45%, on large software projects at 66%, and found that 17% of projects overran by more than 200%, badly enough to threaten the organisation itself. A later McKinsey analysis of IT projects across sectors was blunter still, reporting that only one in 200 projects delivered its intended benefits on time and within budget, with projects overall exceeding budgets by 75% and generating 39% less value than predicted.
Figure
How the services half of a CRM budget behaves
- Large IT projects: average cost overrun45%
- Large software projects: average cost overrun66%
- Large IT projects: value shortfall vs forecast56%
- Projects overrunning budget by 200% or more17%
Averages across large IT and software projects; the licence line rarely moves, the services line does.
Source: McKinsey and University of Oxford, study of more than 5,400 IT projects
None of that is an argument against buying software. It is an argument about which number to negotiate. Buyers spend weeks pushing a vendor from £95 to £78 per user per month and sign a services schedule at day rates without modelling how many days it will run to.
What does CRM implementation actually cost for a five-company group?
Take an illustrative group: five trading companies under one holding company, 40 CRM users spread unevenly across them, one shared account base where three of the brands sell to the same large customers. The cost breakdown over three years:
| Cost line | Annual cost | Three-year cost | What it covers |
|---|---|---|---|
| Per-user licences (40 users @ £100/user/month) | £48,000 | £144,000 | The only line with a published price |
| Implementation and configuration (2x licences) | £96,000 | £96,000 | Permissions, data model, migration, training |
| Integration to finance, ERP, marketing | £24,000 to £48,000 | £24,000 to £48,000 | Per entity, per system |
| Custom reporting for group consolidation | £12,000 to £24,000 | £36,000 to £72,000 | Commissioned per board question |
| Change requests post go-live | £15,000 to £30,000 | £45,000 to £90,000 | At £800 to £1,500 per day |
| Internal admin time (sales ops, team leads) | Not budgeted | Not budgeted | 60% of each user's time per Salesforce |
| Realistic three-year total | , | £345,000 to £450,000 | Licences are 32% to 42% of total |
Assuming an all-in licence cost of £100 per user per month gives £48,000 a year and £144,000 over three years, the number that appears in the board paper.
The services multiple determines where the cost lands. Model the sensitivity:
| Services assumption | Implementation and change spend | Three-year total | Licences as share of total |
|---|---|---|---|
| 1x first-year licences | £48,000 | £192,000 | 75% |
| 2x first-year licences | £96,000 | £240,000 | 60% |
| 3x first-year licences | £144,000 | £288,000 | 50% |
| 3x, plus McKinsey's 45% average overrun | £208,800 | £352,800 | 41% |
The swing between the top row and the bottom row is £160,800, which is more than the entire three-year licence commitment the board actually debated. Ten pounds a seat off the licence price saves £14,400 over the same period. An FD who spends the negotiation on the licence line is optimising the smallest variable in the model. This is the same asymmetry that makes per-seat CRM pricing a poor fit for collaborative groups: the visible meter measures headcount, while the cost is driven by structural complexity.
Why do groups of companies pay more per seat than single companies?
Groups pay a premium because mainstream CRM platforms assume a single trading entity, so everything that departs from that assumption is priced as bespoke configuration work.
A single trading company buys a product that was designed with a single trading company in mind. A group buys the same product and then pays to make it behave like several companies that occasionally act as one. The premium shows up in four predictable places.
First, permissions. Fine-grained visibility control within a brand and between brands is a foundational primitive, and most platforms treat it as a configuration exercise layered on top of a single-tenant data model. Within one brand it is handled adequately. Between brands it turns into custom sharing rules, hierarchies and record-level logic that has to be maintained by someone, forever. That maintenance is a recurring cost, not a project cost, and it is the reason CRM permissions across group companies is the line item most likely to be underestimated at signature.
Second, consolidation. Group-level pipeline and revenue reporting requires a data model that knows an entity exists. Where it does not, the reporting is rebuilt as bespoke work each time the question changes, which is the argument set out in more detail on the cost of custom CRM reports.
Third, corporate change. Gartner's research on M&A-driven software negotiations notes that excessive licence fees arise when contracts define only affiliate use rights without setting out the cost and process for onboarding new affiliates into the existing agreement. For an acquisitive group, that clause is worth more than the discount.
Fourth, the roadmap. Multi-entity groups are a smaller addressable market with a longer, more consultative buying process, so the fundamentals they need rarely justify roadmap space against features that serve the mass market. In practice, that is what pushes a group's requirements out of the product and into the implementation invoice. Those requirements were never bolt-ons; they are architecture, which is why serving them through configuration costs so much.
The structural issue is that multi-entity CRM requirements live in the data model, and no amount of configuration budget converts a single-entity architecture into group-native infrastructure. A group either keeps paying for the gap or removes it.
Figure
The same CRM contract, two different cost profiles
| Dimension | Single trading company | Group of five companies |
|---|---|---|
| Adding a new business | Not applicable | Contract renegotiation plus reconfiguration |
| Cross-entity visibility | Role hierarchy | Custom sharing logic, maintained indefinitely |
| Board-level pipeline view | Standard dashboard | Commissioned report per question |
| Renewal leverage | One user count | Split across entities and budgets |
Illustrative comparison of where cost accrues; Gartner establishes the affiliate-onboarding exposure.
Source: Gartner, Save Millions in License Costs by Preparing for M&A-Driven Software Negotiations
Which costs appear after go-live?
The post-launch leak has two forms, and Gartner has put numbers on both sides of it. Gartner's analysis of SaaS expenditure found that around 25% of software spend is underutilised or overdeployed even as annual SaaS spending grows by more than 19%, and separately concluded that organisations can cut software spending by as much as 30% through licence optimisation, recycling and configuration discipline.
Figure
The recoverable portion of software spend
25%
of SaaS spend underutilised or overdeployed
Gartner analysis of annual SaaS expenditure
up to 30%
of software spend recoverable
Gartner, through licence optimisation and recycling
60%
of a rep's time spent on non-selling tasks
Salesforce State of Sales, seventh edition
The third figure is the one that belongs in a cost of ownership model and almost never is. Salesforce's State of Sales, seventh edition, reports that sales reps spend 60% of their time on non-selling tasks including manual data entry and switching between tools, and that 84% of teams without an all-in-one platform intend to consolidate. On a 40-person commercial team, administrative drag of that scale is a larger annual number than the licence line by an order of magnitude, and it is carried in the sales P&L where no one audits it as software cost.
Most CRM churn arrives dressed as one of two complaints: the system went unused, or it became too complicated and too expensive. Both are the same failure viewed from different ends of the budget. A generic platform either gets abandoned because it fits no one, or gets forced to fit at considerable cost, which is the pattern behind most CRM adoption failure.
What should an FD actually price before signing?
The honest position is that the licence line is the least informative number in a CRM decision, and the most negotiated. The informative numbers are the services multiple, the cost of the next entity, and the standing rate at which the group will commission changes and reports it cannot produce itself. A group that prices those three and finds them acceptable has made a real decision. A group that prices only seats has bought an option on an unknown liability.
Technology cannot fix the underlying issue on its own. If a group's commercial model genuinely requires shared accounts across brands with strict separation elsewhere, that requirement has to live in the data model, and no amount of configuration budget converts a single-entity architecture into a multi-entity CRM. The choice worth making explicit at board level is whether to keep paying for the gap or to remove it.
Frequently asked questions
What is included in CRM total cost of ownership?
CRM total cost of ownership covers per-user licences, implementation and configuration services, data migration, integrations to finance and marketing systems, post-launch change requests, internal staff time spent administering the system, renewal uplifts, and eventual exit or migration costs. Gartner's guidance on estimating CRM project cost advises modelling consulting spend as a multiple of licence fees, which indicates that the services component typically rivals or exceeds the software component.
How much do CRM implementations typically overrun?
McKinsey's research with the University of Oxford, covering more than 5,400 IT projects, found large IT projects run 45% over budget on average and large software projects 66% over, with 17% overrunning by more than 200%. A later McKinsey analysis of IT projects across sectors found that only one in 200 delivered intended benefits on time and within budget. These are averages across project types rather than CRM-specific figures, so treat them as a risk range rather than a forecast.
Why is CRM more expensive for a group of companies than for a single business?
Groups pay a premium because mainstream CRM platforms are designed around a single trading entity, so cross-entity permissions, consolidated reporting and entity additions are delivered as bespoke configuration rather than as product features. Gartner also notes that software contracts defining only affiliate use rights, without a defined process and cost for onboarding new affiliates, expose acquisitive groups to costly renegotiation.
Can any of CRM spend be recovered once the system is live?
Gartner has found that roughly 25% of SaaS expenditure is underutilised or overdeployed, and that organisations can cut software spending by as much as 30% through licence optimisation, licence recycling and configuration review. Recovery on the services and change-request side is harder, because that spend is consumed as hours rather than held as an asset.
Is per-seat pricing the main driver of CRM cost?
Per-seat pricing determines the visible, predictable portion of the bill and rarely the largest portion. Salesforce's State of Sales, seventh edition, reports that 76% of sales leaders say usage-based pricing matters more to customers than a year earlier, though established vendors face a structural obstacle in moving away from seat-based models, since the revenue and renewal base is built on them.
What's the typical implementation cost when deploying a CRM across multiple trading entities within a holding company?
Implementation costs for multi-entity groups typically run at 1x to 3x first-year licence fees. A five-company group paying £48,000 annually in licences should budget £96,000 to £208,800 for initial deployment, covering configuration, permissions logic, data migration, integrations and group reporting, with post-launch change requests accruing at £800 to £1,500 per consultant day.
