A mid-market ERP implementation costs $150K to $750K in the first year and takes 6 to 12 months. The number that surprises buyers is the split: the software subscription is only 20% to 30% of the total. Everything else (the implementation partner, data migration, integration, training, change management) is the bigger bill, and it’s the part nobody quotes in the sales demo.
The odds are the other thing worth knowing before you sign. Gartner expects roughly 70% of ERP implementations over the next three years to miss their original objectives, and Panorama’s 2025 data shows 51% run over budget. ERP is one of the highest-failure-rate projects a mid-market company will ever run. Going in with realistic numbers is most of the defense.
We’re gmware, a software development firm at 5900 Balcones Drive in Austin, TX, with engineering centers in Bangalore and Mohali, India. We don’t sell you an ERP. We build the integration layer that makes one actually work inside your business, and we’ve watched enough of these projects up close to write down where the money goes, where the timeline slips, and the one decision that separates the projects that land from the ones that become a cautionary slide.
ERP implementation, in four numbers
What does an ERP implementation cost in 2026?
It scales with headcount, and the bands are wide because scope is wide. A small business at $25K to $150K is usually buying a cloud platform with light configuration. A mid-market company (100 to 500 employees) lands at $150K to $750K. The upper mid-market runs $500K to $2M, and a true enterprise rollout starts at $1.5M and climbs past $5M. As a sanity check, companies tend to spend 1% to 3% of annual revenue on the project.
| Company size | Typical first-year cost | Typical timeline |
|---|---|---|
| Small business (under 100 employees) | $25K to $150K | 3 to 6 months |
| Mid-market (100 to 500 employees) | $150K to $750K | 6 to 12 months |
| Upper mid-market (500 to 2,000 employees) | $500K to $2M | 9 to 18 months |
| Enterprise (2,000+ employees) | $1.5M to $5M+ | 12 to 24+ months |
That first-year total bundles the subscription, the implementation partner, data migration, integration, and training into one figure. The annual subscription by itself is much smaller: cloud ERP runs $50 to $200 per user per month, with NetSuite around $125 and SAP S/4HANA around $200 per user. Multiply that out and you get a number that looks affordable. Then the implementation quote arrives, and it’s two or three times the software.
Why is the software the smallest line on the invoice?
Because the platform is the easy part. The vendor already built the accounting engine, the inventory logic, the procurement workflows. What costs money is bending that platform around how your business actually runs, moving twenty years of data into it, and wiring it to every other system you own.
The split is consistent across the market. The license or subscription is 20% to 30% of the total. The implementation partner’s fees alone run 40% to 60% of the build, which is why implementation services typically cost 1x to 3x the first-year software fee. At $150 to $350 an hour for North American ERP consultants, those hours add up fast.
Where the ERP budget actually goes
The six cost drivers, in plain terms
Strip the proposal down and a quote is built from six things. The subscription or license. The implementation partner’s labor. Data migration, which is harder than anyone wants it to be. Integration to your other systems, priced per connection. Training. And change management, the line that’s always too thin.
A few of these have their own published ranges worth pinning down. Data migration for a mid-market project runs $15K to $150K+ depending on how messy the source data is. Each system integration runs $5K to $50K. Customization is the silent multiplier: a moderate level of it adds 25% to 50% of the base license cost, and heavy customization 50% to 200%. And training plus change management, the two things adoption actually depends on, get funded at 5% to 10% of the budget when they should get 15% to 20%. That underfunding shows up later as a system nobody uses.
How long does an ERP implementation take, and where does it slip?
A mid-market rollout takes 6 to 12 months. Small-business core go-lives run 3 to 6 months; multi-entity enterprise programs run 12 to 24 months or more. Those are the honest planning numbers. The dishonest ones are whatever the timeline says when no one has looked at your data yet.
The phases are predictable. Discovery and planning come first, then process design and configuration (the longest phase, and where scope creep does its damage), then data migration, then testing and training, then go-live and a stabilization period that’s longer than anyone budgets. Most of the slippage hits in design and migration, not at go-live.
Here’s where it actually slips. 64% of ERP projects exceed their original timeline, and data-quality problems alone account for 41% of those overruns. Customization piles on more: it can add 25% to 30% to the original schedule. The pattern is the same every time. The plan assumed clean data and a stable scope. The project found dirty data and a moving scope. The two-week migration becomes two months while the team cleans records that should have been audited before kickoff.
Why do roughly 70% of ERP implementations fail?
Gartner expects about 70% of ERP implementations over the next three years to miss their original objectives, and its longer-running figure puts 55% to 75% of ERP projects falling short of their goals. Those are brutal odds for a project this expensive. The reassuring part, if there is one: the failures are operational, not technical. The software almost never breaks. The project around it does.
Panorama’s 2025 data names the budget side precisely. Of the 51% that run over budget, the leading causes are underestimating project staffing (38%), expanding the initial scope (35%), and technical or data issues (34%). Notice what’s not on that list: the vendor’s product. People underestimate their own people, let scope drift, and discover their data is a mess. None of that is the ERP’s fault.
Why budgets blow (Panorama 2025)
The ERP failure checklist, with the fix for each
Every failure mode we’ve seen has a counter, and every counter happens before configuration starts. Run this list against your project. If you can’t check a box, that’s the work to do first.
| Failure mode | What it looks like | The fix |
|---|---|---|
| Underestimated staffing | Your team is doing the project on top of their day jobs, and it shows by month three | Backfill or budget real internal hours; underestimating staffing is the #1 overrun cause at 38% |
| Scope creep | ”While we’re in here” requests pile up; the go-live date drifts | Freeze scope after design; route new requests to a phase two with its own budget |
| Dirty data | The migration stalls while the team cleans records mid-project | Audit and clean data before kickoff; data quality drives 41% of timeline overruns |
| Over-customization | The platform gets bent so far it’s expensive to maintain and upgrade | Change the process to fit the ERP where you can; customize only the genuine differentiators |
| No executive owner | A committee sponsors it, so no one makes the hard calls | Name one accountable owner with budget authority, not a steering committee |
| Underfunded training | The system launches, adoption stalls, people route around it | Fund training at 15% to 20%, not the typical 5% to 10% |
| No integration plan | The ERP can’t talk to your other systems; staff re-key data by hand | Scope the integration layer as a first-class line item, covered next |
The cost of skipping this is concrete. Panorama documents a food-and-beverage company that lacked project governance, hit serious scope creep, and ended up with a 25% cost increase and a six-month go-live delay as a result. That’s not an exotic outcome. It’s the median outcome when these boxes go unchecked.
Why does an off-the-shelf ERP still need custom integration?
Because it ships as an island. NetSuite, Dynamics, and SAP are excellent at what they do, but out of the box they don’t talk to your CRM, your ecommerce store, your warehouse management system, your shipping provider, or the custom app your operations team lives in. Connecting them is real engineering, and it’s the line most buyers discover late.
Each integration runs $5K to $50K. A CRM connection lands around $5K to $25K, an ecommerce integration $8K to $50K, and the price climbs with data volume, custom fields, and how much real-time syncing you need. A business with six surrounding systems is looking at a meaningful integration program, not a checkbox.
This is the part that doesn’t fit the ERP vendor’s job description, and it’s exactly the part we do. An ERP that can’t exchange data with the systems around it forces your staff to re-key everything by hand, which quietly cancels the efficiency you bought the ERP for. The integration layer is what turns a system of record into a system that actually runs operations. We treat it as its own scoped project with its own budget, because pretending it’s free is how a clean ERP rollout turns into a manual-data-entry job. If your real problem is that your systems don’t talk to each other, our operations and process management work starts there.
When does an off-the-shelf ERP beat a custom build?
Almost always, and we’ll defend that even though we’re a custom software shop. Here’s the honest version.
Buy the ERP. The accounting, inventory, procurement, and financial-reporting logic inside NetSuite or Dynamics took those vendors decades and thousands of customers to get right, including the tax rules, the audit trails, and the compliance edge cases you don’t want to discover you missed. Rebuilding general ledger and inventory management from scratch is a multi-year project that, on the day it ships, does roughly what a $125-per-user subscription already does. That math almost never works. We’ve talked more than one company out of it.
What you build is the layer the ERP can’t give you: the integrations to your other systems, and the genuinely unique workflow that is your competitive edge and doesn’t exist as a configurable option in any platform. That’s the right place for custom engineering. It’s also a fraction of the spend and risk of a full custom build, and it pays back because it’s solving a problem no off-the-shelf product can.
The exception that proves it: if your core operation is so specific that no ERP on the market models it, and that operation IS your business, a custom operational system can be the right call. That’s rare, and it’s a different project from “we want an ERP.” We dug into that build-versus-buy line for smaller teams in our guide to custom software development cost for small business, and the same logic scales up: build the differentiator, buy the commodity.
| Off-the-shelf ERP + custom integration | Fully custom build | |
|---|---|---|
| What you get | Proven accounting, inventory, procurement, plus connected systems | A system built exactly to your processes |
| First-year cost | $150K to $750K mid-market | Far higher, with a multi-year runway |
| Risk | Lower; the core logic is battle-tested | Higher; you own every edge case and bug |
| Best for | The vast majority of mid-market companies | A truly unique core operation that is the business |
Buy the commodity, build the difference
How gmware approaches an ERP project
We’re not an ERP reseller, and that shapes how we work. We don’t earn a license commission, so we have no reason to push you toward a bigger platform than you need or to bill customization hours bending it into knots. Our job starts where the platform stops: the integration layer that connects your ERP to the systems around it, and the data work that decides whether the migration is two weeks or two months.
Before we quote, we audit the data and inventory the systems that need to connect, because both change the number, and a quote built without them is a guess. That discipline isn’t theory for us. We operate Shield Suite, a retail-intelligence product that ingests data across more than 60,000 beverage-alcohol storefronts, so integrating messy data from many systems is a thing we do to ourselves daily, not a slide. Architecture and client-facing leadership sit in Austin; build capacity runs from Bangalore and Mohali with overlapping US hours, which keeps senior oversight on your timezone without US-only rates. When the ERP is one piece of a wider operational overhaul, a digital transformation program wraps around it, and we’ll scope that honestly too, the same way we break down the rewrite-versus-refactor math on legacy systems.
Tell us which ERP you’re evaluating and what it needs to connect to. Reach out and we’ll give you a straight answer on integration scope, cost, and timeline within 48 hours. If the honest answer is that your current systems just need to talk to each other and you don’t need a new ERP at all, we’ll tell you that too.