Set the ERP budget from the published ratios, not the number on the vendor's quote. Implementation runs 1x to 3x your first-year software fee, the systems integrator takes 40 to 60% of that, and the system costs 15 to 25% of the implementation figure every year to run, by ERP Research, an independent research firm. Six line items decide whether you land low or high in the range. Your own people add 15 to 30% on top, and two contingencies, one for cost and one for schedule, sit above that. Present the five-year total that results, not the go-live figure, or the board judges a success as a failure.
Build the budget from two ratios and the vendor's annual price
A defensible budget starts with one number you already have: the annual software price on the vendor's quote. Call it S, the subscription or the maintenance bill plus the amortized licence. Two multiples turn S into a program budget.
The first is implementation to first-year software. ERP Research puts services at 1x to 3x the first-year fee, and within that the SI takes 40 to 60%. The second is the five-year total to software, which the same source puts at 3x to 5x. That second rule was built for a one-time licence. Under a subscription the five-year total runs higher, because five years of subscription alone is already 5x the first-year fee before implementation or run cost is added.
Run cost closes the model. The system costs 15 to 25% of the implementation figure every year after go-live, per Ecosire and SysGenPro. The table below collects the ratios this article uses. Each is published by a research firm or a cost benchmarking site, and each carries a confidence grade. B means two or more independent sources agree. C means a single interested source, attributed and treated as a planning range.
| Ratio | Published value | Source | Confidence |
|---|---|---|---|
| Implementation services to first-year software | 1x to 3x | ERP Research | B |
| Five-year total cost to software | 3x to 5x | ERP Research | B |
| SI fees as share of implementation | 40 to 60% | ERP Research | B |
| Annual run cost after go-live | 15 to 25% of implementation | Ecosire | C |
| Internal team, backfill, and change | 15 to 30% on top of external spend | ERP Research | B |
Work it through on a quote. ERP Research lists Oracle NetSuite at about $999 a month plus $99 per user. For 300 users that is 300 times $99 times 12, or $356,400, plus $11,988 base, so S is about $368,000 a year before discount. Implementation at 1x to 3x is $368,000 to $1.1M. The SI inside that is $147,000 to $662,000. Those two lines of arithmetic give you a budget range before a single SI has scoped the work, which means you can test the proposal rather than accept it.
Six line items decide where you land in the range
The ratios set the width of the range. Six line items decide where inside it you land, and five of the six are settled before the SI contract is signed. Walking each one against a written list keeps them from hiding inside the SI's single number; the ERP budget checklist sets them out in the order they bite.
| Line item | Published range or swing | Control lever | Source (confidence) |
|---|---|---|---|
| Custom code | Swing of 20 to 40% of implementation cost | Cap custom code at 10 to 15% of requirements; make it a steering-committee rule | ERP Research (B) |
| Data cleanup | 10 to 15% of the program budget | Start cleanup before design, with named data owners on your payroll | Costbench (C) |
| Integrations | $10,000 to $50,000 each | Cut the interface count before build | ERP Research (B) |
| Dual-run | 12 to 24 months of paying for both systems | Fix the cutover date; shorten the parallel period | Costbench (B/C) |
| External program office | $5,000 to $20,000 a month | Hold the timeline; move the role in-house after go-live | ERP Research (B) |
| Rescue reserve | 30 to 50% of the original budget | Commission an independent review at the first warning signs | Deliberate Directions (C) |
Custom code is the largest lever a buyer controls. ERP Research finds that holding it under 10 to 15% of requirements saves 20 to 40% of implementation cost, and the finding holds across tiers, because custom code costs once to build and again at every upgrade. On a $740,000 implementation, that swing is $148,000 to $296,000. Data cleanup is second, at 10 to 15% of the program budget per Costbench, and it lands on your own finance and supply-chain staff, not the SI's, so it rarely shows in the quote.
Integrations run $10,000 to $50,000 each, so a company with 12 interfaces adds $120,000 to $600,000, and the count is the number to watch. Dual-run means paying for both the old and new systems for 12 to 24 months. The external program office runs $5,000 to $20,000 a month, which over a 36-month tier 1 program is $180,000 to $720,000. The sixth item, a rescue reserve, is the one no approval paper includes. Deliberate Directions, an advisory firm writing about NetSuite rescues, puts a rescue at 30 to 50% of the original budget. That is a single interested source on one product line, so treat it as the tail you guard against, not a line you fully fund. The six line items that blow up ERP budgets takes each apart with its own worked arithmetic.
Add your own people at 15 to 30% on top of the external spend
The ratios and the six line items price what you pay outside the company. They leave out the most underbudgeted cost in any program, which is your own people. The internal team, backfill for the business leads pulled onto the program, and change management add 15 to 30% on top of the external spend, per ERP Research's hidden-cost estimate.
The reason this is missed is that it never appears on an invoice. The SI quotes its own hours. Your finance controller who spends half her year on the program, the warehouse supervisor who leaves the floor to design the new process, and the backfill hired to cover both, all sit on your payroll and your budget, not the SI's.
Put numbers on the NetSuite example. Take implementation at the 2x midpoint, about $740,000. The internal uplift at 15 to 30% is $111,000 to $222,000, and it runs for the length of the program, not just at go-live. On a tier 1 program at $9M of external cost, the same uplift is $1.35M to $2.7M. A budget that omits this line is not conservative. It is wrong by about a fifth, and the gap surfaces as a staffing crisis in the middle of the program when the day jobs still have to be done.
Price two contingencies, one for cost and one for schedule
A budget built from ratios and line items is still a point estimate, and ERP programs miss point estimates. A survey reported by ERP Today found 31% of implementations took longer than expected and 12% had significant delays. Two separate contingencies cover the two ways a budget breaks.
The first is a cost contingency, for scope found late and change orders on the six line items. The largest single swing is custom code at 20 to 40% of implementation, so a contingency that cannot absorb at least part of that is not sized to the risk. The rule the site recommends is a cost contingency of 10 to 15% of the external total, held centrally by the steering committee rather than spread across the plan where it gets spent early. That is the site's judgment, not a published figure.
The second is a schedule contingency, and it is priced differently: monthly burn times months of slip. Every month a program runs long has a fixed cost that keeps meters running even when the build is stalled. Consider a $500M manufacturer with five plants running a tier 1 program, with a six-month dual-run planned. During a slip it pays three things at once. The new subscription, say $1M a year, is $83,000 a month. Maintenance on the old system, at SAP's 22% of licence value on a $2M licence base per ERP Research, is $440,000 a year, or $37,000 a month. The external program office at the top of the range is $20,000 a month.
That is about $140,000 a month. A six-month slip, which sits inside the published delay rates, is $840,000 before the SI sends a single change order. Nothing in that example happened to any company. It is arithmetic on published figures, and it is why the schedule contingency belongs in the approval paper as its own line: the planned monthly burn through a slip, multiplied by the months of slip you are willing to fund. A board that sees both contingencies, sized and separated, can tell the difference between a program that is over budget and one that is simply later than hoped.
Present the five-year total to the board, not the go-live number
The board remembers the number you give it at approval and measures every later invoice against it. If you present the go-live cost, the subscription renewals, the run cost, and the internal team all arrive afterward and each reads as an overrun, even when it was always going to be spent. Present the five-year total instead. Written for a five-year horizon, the external cost is:
Five-year external cost = 5S + mS + 4r(mS)where m is the implementation multiple of 1 to 3, and r is the annual run cost share of 0.15 to 0.25. The multiplier on run cost is 4, not 5, because support in the first weeks after go-live usually sits inside the SI's implementation fee. Add dual-run and the 15 to 30% internal uplift as their own lines, because neither scales with a software discount.
Run the NetSuite example end to end. Five years of software is $1.84M. Implementation at 1x to 3x is $368,000 to $1.1M. Run cost for years two through five is $221,000 to $1.1M. The external total is $2.4M to $4.05M, and the internal uplift brings the fully loaded five-year total to about $2.8M to $5.3M. That is 7.6x to 14x the first-year software fee, which is why a five-year number quoted at the old 3x to 5x rule has quietly dropped four years of the subscription.
The same build at tier 1 scale, using Dynamics 365 Finance at its $210 per user per month list price for 1,500 users, gives S of about $3.78M and a fully loaded five-year total near $29M to $54M. That lands inside the $20M to $50M range published for a global tier 1 program, which is the check that the model is calibrated. A five-year ERP TCO model the board can defend shows the full build on one slide. The figures above are list prices for illustration; tier 1 contracts mix full and lighter user types and every line scales with a negotiated discount, so quote the total at both list and the negotiated price so the board sees what the negotiation saved.