Most IT leaders can quote their ServiceNow licence cost in a heartbeat. Almost none can tell you what it actually costs to run the service management capability end to end, or which parts of that cost they can genuinely influence.
Ask a CIO what their ServiceNow contract costs and you'll get an answer in seconds. Ask what it costs to process the work flowing through that platform, and the room goes quiet.
That gap matters, and it isn't just an IT question. Licence negotiations can save real money. So can rightsizing the internal team running the platform, reducing reliance on external consultants, and improving how efficiently work actually gets handled once it's in the system. All four levers are real. Most organisations can only see one of them.
Platforms like ServiceNow can hold enormous numbers of records. Most of them aren't useful units for a cost conversation. Events, alerts, notifications, access control rules, configuration items and assets are essential to running the platform, but they aren't the work people are actually trying to get done.
The unit that matters is the process record: an incident, a request, a change, an HR case, a customer case, a security case, a risk, an audit finding, a facilities request, anything that represents a person or team actively moving work forward.
Once process records become the unit of measurement, a much sharper question becomes possible: what does each one actually cost the business? We call this the now2value ESM Efficiency Model. ServiceNow is our own frame of reference throughout this piece, but the model itself isn't ServiceNow-specific — it applies to any company running any enterprise or IT service management platform.
How efficiently do we process the work that enters our service management platform, and where does automation actually move the number?
PR = Process Record — a unit of actual work: an incident, request, change, HR case, customer case, security case, risk or audit finding. Not events, alerts, configuration items or other foundation records.
Dividing by PR turns raw totals into a comparable rate — the only way to benchmark cost or efficiency across companies, business units or years, regardless of size.
Platform cost covers three things, all real levers on their own: licence and subscription spend, the internal team that operates and develops the platform, and external consultancy or managed-service spend. Processing cost is driven by how much active time it takes a person to move a record through its process, and how much of that work still requires a person at all.
Take a company with 10,000 employees running a broad service management deployment in ServiceNow, covering IT, HR, facilities and more.
The €3.1M platform cost splits across a €1.8M licence for a multi-module deployment, a €0.9M internal platform team (a blend of Denmark-based and near-shore roles), and €0.4M in ongoing external consultancy. Add the human side: handling time averages 15 minutes at a blended labour rate of €45/hour, with a quarter of records genuinely touchless. Platform spend alone would never have surfaced the €23.95 figure. It takes both halves of the model to see it.
The same model makes automation tangible rather than aspirational. If touchless automation moves from 25% to 35%, with volumes, handling time and labour rates unchanged, processing cost drops by roughly €225,000 a year.
Moving that percentage point is rarely about typing faster. In practice, it's AI doing the work: GenAI drafting responses to high-volume, low-variation requests, agentic workflows completing multi-step tasks without a person routing each one, and virtual agents deflecting requests that never needed a human at all. The automation rate in this model is where an AI investment shows up as a number, not just a capability.
Worth being precise about what that number actually is: it's released capacity first, not cash in the bank. It might show up as faster service, more resilience, or room to absorb growth without adding headcount, before it ever shows up as a smaller invoice. But it gives leaders a shared financial language for deciding where automation investment actually pays off, instead of arguing about it in the abstract.
Processing Cost expands once automation enters the picture — this is the lever the calculator below lets you move.
Move Automation Rate toward 100% and Processing Cost falls toward zero — the calculator below shows exactly how much, in your own numbers.
None of these need to be precise on day one. Start with ranges, then refine with platform telemetry and targeted time studies.
A cost-per-record number shouldn't turn into a crude benchmark against other companies. Record mix matters enormously — a password reset and a major incident aren't economically equivalent, and geography, labour model and service complexity all shift the number. The value isn't the single figure. It's decomposing it: is cost high because the platform team is oversized? Because customisation has created ongoing maintenance overhead? Because external consultancy has become structural rather than transitional? Because too many records still move through email and spreadsheets instead of the platform? Or because automation is genuinely low?
That's also why a genuine benchmark — what good Cost per Process Record actually looks like for a 5,000-, 10,000- or 50,000-employee enterprise — has to control for exactly these variables to mean anything. It's a real next step, not something one worked example can responsibly claim.
There's a question this model deliberately doesn't answer. Picture two comparable 10,000-employee companies in the same industry. One generates 1,000 process records a day, the other 2,000. The first might simply run more reliable services, with simpler processes and less operational friction — in which case, lower volume is a real quality signal.
But lower volume isn't automatically better. It can just as easily mean poor ESM adoption, unreported issues, or work quietly moving through email and spreadsheets instead of the platform. Folding record-generation quality into a unit-cost formula would make the number harder to interpret, not more useful.
Cost per Process Record measures the economic efficiency of operating the ESM capability and processing the work captured within it. It does not assess whether the organisation generates an appropriate number of process records. Record generation may be influenced by service quality, process maturity, organisational complexity, ESM adoption and the scope of processes managed on the platform.
The now2value ESM Efficiency Model answers one question well: how economically does the organisation operate the platform and process the work captured within it? Whether it should be generating more or fewer records in the first place is a separate quality and demand question — deliberately kept out of this formula, not ignored.
There's always a temptation to make an efficiency model more complete by adding more variables. The risk is that a genuinely useful metric turns into a complicated maturity model instead. Seven inputs and one clearly defined output is what makes Cost per Process Record easy to explain, benchmark and actually use in a real executive conversation — not a limitation to apologise for.
For CIOs, CTOs and CFOs, this turns service management from a technology-cost line into an operating-model question: how efficiently does the organisation convert a platform, its people and its automation into completed work?
Once that question can be expressed as platform cost per record, processing cost per record, and the size of the improvement opportunity, decisions get easier to compare — whether the right move is licence renegotiation, rightsizing the platform team, reducing consultancy dependency, process redesign, or automation. None of these levers is automatically the answer. The model's job is to show you which one actually is, for your organisation, this year.
Most organisations have never put all four levers on the same page. Let's find out together.
Speak with us