Data conversion is one of the most technically demanding parts of any ERP implementation—and one of the most frequently underestimated.
By the time most teams start thinking seriously about data quality, the window to fix it affordably has already closed. Old records have accumulated. Legacy structures don’t map cleanly to the new system. And pressure to stay on schedule makes it harder to stop and clean up the mess that’s been years in the making.
RPI’s Supply Chain Practice Manager Dan Farruggio will walk through what a proactive data conversion strategy actually looks like and what you can do right now to reduce cost and risk.
Session attendees will learn how to:
- Identify data quality risks early enough to address them before they become costly
- Build a conversion strategy that reduces rework and schedule pressure at go-live
- Monitor the right indicators during active conversion to catch problems before they compound
- Validate data integrity post-conversion with confidence
- Establish practices that keep data clean and reliable long after go-live
No matter what stage of your ERP project your business is in, this session offers valuable perspective on how to handle data conversion.
Transcript
Dan Farruggio
Hello everybody — good morning or afternoon, wherever you’re joining from, and welcome. I’m Dan Farruggio, practice manager for the supply chain team here at RPI Consultants, and I’m genuinely glad you’ve made time for us today. Today we’re going to be talking about data conversions — one of those topics that every organization knows matters and almost every organization underestimates. After 14 years in this space, I can tell you without hesitation: organizations that treat data as a strategic asset before their ERP goes live are the ones that actually realize the return on their investment. These are the ones that don’t spend the first year firefighting issues and complaints. So today we’re going to have a real practitioner-level conversation about what separates organizations that nail their data conversion from those still cleaning up 12 months post-go-live. Let’s get into it.
Here’s how we’ve structured the next 25 to 30 minutes. Every section is grounded in what we see in the field today with healthcare and public sector clients. We’ll start by making the case that successful data conversions begin long before the implementation even starts. Then we’ll get practical with actions you can put in place right now. From there, we’ll cover the critical watchouts during your actual conversion process. And finally, how do you sustain data quality after go-live? Because that post-conversion window is where value gets captured — or quietly erodes.
A quick background on me before we dive in: I’ve been with RPI Consultants since 2011, coming up on 15 years, focused almost exclusively in supply chain with some security and process automation expertise along the way. Over those years, I’ve worked with complex healthcare systems and public sector organizations, and I’ve found that technical implementation is rarely where these projects struggle. The real challenges are almost always people, process, and data — and of those three, data has the longest tail.
I’m based in Southwest Florida, where I live with my wife and three kids. When I’m not working, you’ll find me at the beach, on the pickleball court, or anywhere outside with my family. And if you ask me why RPI — it’s the people, hands down. I love working with a talented group of professionals who care about outcomes and who understand they can’t be successful until you are. You’ll see that reflected in everything we discuss today.
All right — it’s never too early to start. The success of your conversion begins before your implementation even kicks off.
More than half of the issues we see throughout an ERP implementation can almost directly be traced back to the quality of legacy data. It’s not the system configuration, it’s not the integrations — it is the data that was brought over from the old system. The implications of that bad data touch every dimension of your supply chain operation: inventory accuracy, vendor payment integrity, contract pricing, audit outcomes, and whether your end users actually trust the system they’re being asked to adopt. If they don’t trust the data, adoption suffers — and users usually start breaking policy and procedure just to get their jobs done.
Think about the promise of any modern ERP platform: the analytics, the dashboards, the real-time visibility. Those capabilities are only as good as the data supporting them. A beautiful reporting tool or an amazing dashboard built on dirty data is just a more elegant way to make a bad decision. The organizations that get this right don’t start thinking about data six months before go-live. They start two or three years out. They treat data readiness as an ongoing discipline, not a project task. That investment pays dividends far beyond the implementation itself. So ask yourself: is data quality in your organization something you address reactively, or is it built into how you operate on a daily basis today?
How you prioritize your data work depends on your industry. I’ll spend a minute on the two verticals where we do the majority of our supply chain work. In healthcare, primary concerns tend to be item master complexity — clinical versus non-clinical items and locations — unit of measure inconsistencies, contract pricing accuracy that flows into patient billing, and inventory data feeding downstream analytics and charge capture. Errors in these areas compound fast. On the public sector side, priorities often shift to fund and grant management integrity, encumbrance accuracy, audit trail requirements — which are typically more stringent than in the private sector — and the challenge of consolidating a decentralized buying structure where departments have historically done things their own way.
Frame your data readiness work through this lens: what are the high-stakes errors you cannot afford to carry into your new system? Start there.
When most people hear ‘supply chain data conversion,’ they automatically think items and vendors. That’s the core, but it’s just the beginning. We break supply chain data into three domains. Master data — your items, vendors, and contracts — the foundational records that everything else references. Transactional data — your open purchase orders, pending receipts, and inventory balances — your active operational state. And reference and control data — your movement classes, report groups, job screens, and security configurations. This last category is often overlooked or addressed late in a project, and it’s frequently where post-go-live surprises originate. Make sure you spend time here.
Here’s a diagnostic question I like to ask clients: do you know where your data comes from and where it goes? Which systems feed your ERP, and which systems consume from it? A clean conversion isn’t just about what’s inside the system — it’s about what flows in and out every day. The stakeholders impacted by your data quality extend well beyond supply chain. Finance, operations, compliance — they all have a stake in this conversation, and it’s important to include them whenever changes are being made.
Let me make this concrete with two real-world case scenarios — one for healthcare and one for public sector. A large healthcare system was preparing for an ERP upgrade with a general sense that their item master had grown over the years. When we assessed it, they had over 150,000 item records across multiple facilities. After a structured cleansing effort — working with both clinical and supply chain teams to identify what was genuinely active versus obsolete — they landed at around 45,000 records. That’s almost a 70% reduction in conversion complexity. Seventy percent fewer records to map, validate, test, and troubleshoot.
Beyond the volume reduction: vendors were merged, historical transactions were corrected, units of measure were standardized, naming conventions were normalized, and accurate crosswalks were created. After that first pass, the conversion effort was a fraction of what it would have been, and go-live had dramatically fewer data-related incidents. The takeaway: deduplication and normalization done in your legacy system before conversion isn’t cleanup work — it’s a strategic investment that pays off on every conversion pass throughout the project, not just the last one.
The parallel story from the public sector side: a local government heading into an ERP upgrade had a classic decentralized procurement problem. Years of departmental autonomy had produced duplicate vendor records, inconsistent commodity codes, and contract structures that varied widely across the organization. The significant decision leadership made was to fund a standalone data assessment before the project even kicked off. What that assessment uncovered was a little uncomfortable, but necessary — it revealed how years of inconsistent process had compounded into a systemic data quality problem. Because they found it early, there was time to act on it. Thousands of duplicate records were eliminated before the first conversion pass. Vendor reporting improved. Contract compliance tracking became meaningful. And training was conducted on data the team could actually defend and follow. Early identification always beats late discovery. If there’s one thing you take from today, that might be it.
What those case studies illustrate is part of a broader trend we’ve been tracking. Organizations are no longer waiting for an implementation contract to be signed before addressing data quality. Data readiness programs are now being launched 12 to 24 months — or more — ahead of ERP projects kicking off. In some cases, before the ERP selection is even finalized. Before they know which system they’re going to, organizations are standing up initiatives to get their data clean. Governance communities are being formed, data stewards are being assigned, and cleansing initiatives are funded as standalone work streams.
Why this shift? ERP timelines are compressing, particularly with cloud deployments — there’s less margin for error. Vendors are clear that data quality is the client’s responsibility. And organizations that have been through painful go-lives are making a different choice the second time around. They’re learning from the experience and acting sooner. Data cleansing has evolved from a technical task at the bottom of the project plan into a strategic transformation activity. Organizations treating it that way are the ones that are winning.
Here’s a simple framework to use as a foundation. First, establish ownership. Data without owners is data without accountability. Assign business data owners for your key supply chain domains — items, vendors, contracts, inventory locations. These should be functional leaders and subject matter experts, the people who live and breathe this data every day, not just IT resources.
Second, inventory and classify your data. You can’t improve what you can’t see clearly. Even a focused item master assessment can surface significant findings quickly. Third, begin targeted cleansing. Don’t try to boil the ocean. Start with the highest-risk areas. In healthcare, the item master and contract pricing are usually good starting points. In the public sector, vendor records and fund structures tend to be the right place to begin.
Fourth, document your business rules: how data gets created, who approves it, how it gets retired, and what the archive strategy will be post-go-live. These questions are the foundation of your governance model and your conversion scope. There’s no bad time to start thinking about this work, but there absolutely is a point at which it becomes too late to receive the full benefits. Don’t wait — start early.
Here are seven data cleansing activities that consistently deliver the highest return. Inactive items: if something hasn’t moved in 24 to 36 months, it probably doesn’t need to be converted. Carrying it over inflates your conversion complexity and confuses users. Expired contracts: they may still be technically open in the system, but if they’re not active, close them before you migrate. Duplicate vendors: this is the enemy of accurate spend analytics. Go through the effort of merging them with transactional history — it makes life significantly easier downstream, and you’ll be glad you did. Naming conventions and standardization: invisible in summary reports, but devastating when users are trying to search or run matching logic. Make sure you have a standardized naming convention for items and descriptions. Duplicate items: often the result of multiple facilities ordering the same product under different names or vendor relationships. Consolidation here drives real savings and better analytics. Units of measure: the difference between a case and a box can mean ordering ten times what you intended. Spend time going through your units of measure and conversion factors to get them clean. And finally — often overlooked — if you’re an EDI customer, go through your EDI trading partners and old records. Decommission inactive ones, clean your substitution records, validate your active connections. Remove anything that’s no longer needed in your future state.
Once you clean your data in the legacy system, you benefit from that work on every conversion pass going forward. The effort you put in early compounds throughout your entire ERP project.
Here’s another trend worth calling out — one that’s reshaping conversion scope altogether: how much historical data is actually worth converting. For years, the standard expectation was seven to ten years of transactional history. That expectation drove enormous conversion complexity, significant testing effort, and higher cost. What we’re seeing now is a meaningful shift. Organizations are moving to converting open transactions and current balances only, while historical data is archived externally — in data warehouses, data lakes, or read-only legacy environments accessible for reference.
The drivers are practical. Cloud ERP platforms are optimized for current operations, not warehousing decades of history. Implementation timelines don’t accommodate that testing burden. And modern archiving solutions have made the alternative genuinely viable. You don’t have to bring everything with you — you can offload historical data into a warehouse or data lake and access it when needed. This is absolutely a conversation worth having with your leadership and implementation partner before your scope decisions get locked in. The choices you make here have significant downstream implications for your cost, timeline, risk, and overall project complexity.
The conversion process during implementation involves multiple passes. Each pass is an opportunity to test your data mapping, validate your business rules, and resolve any errors that surface. This builds your team’s confidence to execute the final conversion with conviction. The decisions that should be made before the first conversion pass are: how much history are you converting? What standards qualify as clean enough data for the new system? And what stays in your legacy or archive system? These are business decisions that require functional leadership to own — or at minimum, be a meaningful part of.
The two failure patterns I see most consistently: decisions not resolved before testing began, and validation performed by technical teams without meaningful functional engagement — validating by record count instead of actually examining the details. Your procurement leaders and inventory managers need to own validation. They’re the ones who will live with what goes live. Validation is an iterative issue resolution process across every conversion pass, and it’s where you build quality into your process — not just check for it at the end.
I would definitely not recommend cutting corners before go-live. Here are the critical checkpoints. Inventory balances: every unit must reconcile between your legacy and new system. If it’s wrong on day one, you’re starting in a hole that won’t fix itself. Contract pricing: bad pricing creates supplier disputes and, in healthcare, billing compliance risks. Validate before conversion and make it a gate to move forward. Supplier records: vendor addresses, payment terms, banking information — if these are wrong, your first payment run will be a fire drill. Units of measure and conversion factors: make sure they’re accurate across all items. EA versus CS for case, for example — errors here cause real operational damage that takes significant time to unwind. Reporting accuracy: run your standard reports against the converted data well ahead of go-live. If the numbers don’t hold up, you’re not ready. Understand any discrepancies and get as close to 100% accurate as possible. And finally, make sure your job screens and distribution groups are set up correctly. If automated jobs or approval routing don’t function after conversion, everything downstream stalls. Build a formal validation checklist, don’t go live with any unchecked boxes, and make sure your functional teams are the ones validating and signing off — so they own their data.
A few common pitfalls to be aware of — even well-prepared teams can get burned here. Assuming your legacy data is good: we hear this regularly. ‘Our data is pretty good and clean.’ And then, after 10 years of workarounds baked into the records, that turns out not to be the case. Don’t assume — always validate. If the answer to something is ‘this is the way we’ve always done it,’ that’s the first place you should look to understand why and whether it can be done better.
Underestimating mapping complexity: field-to-field mapping is almost never one-to-one. If you skip the deep analysis, you’ll find the gaps at exactly the wrong time — during testing. Don’t be afraid to do a thorough dive and know precisely where your data is going and whether it’s mapped accurately. Leaving decisions until testing: it’s tempting to say ‘we’ll decide that when we get to SIT or user acceptance testing,’ but that’s too late. Unresolved business rules become bottlenecks that compound under the pressure of project sprints, leading to shortcuts and incomplete solutions that carry over into post-go-live problems. Make decisions early and understand their impacts. And finally, relying only on summary validation: record counts confirm that data arrived, but they don’t confirm it arrived correctly. You need record-level validation and scenario-based testing to have genuine confidence in your data.
Now, what do you do once you’re live? The first 30 to 60 days post-go-live are the most critical window in your entire implementation journey. This is when data quality issues that slipped through validation become visible — and when they’re still addressable before they compound. The KPIs I’d encourage every director in this session to monitor personally during that window: inventory reconciliation, PO and contract verification, vendor payment execution, approval workflow validation, EDI and work queue performance, and job and report accuracy.
These are business metrics, not just IT metrics. They warrant executive visibility, daily monitoring, and regular reporting. If anything stands out, escalate it to your team or implementation partner for review. Early detection here is the difference between a manageable speed bump and a project that starts running off the rails. This 30-to-60-day window is also a key period for monitoring system adoption — how your users are taking to the new environment.
Once you’re stabilized — around the 60-to-90-day mark — the conversation shifts to optimization. This is where organizations consistently leave value on the table. Go-live is not the finish line; it’s the starting gun for phase two and the rest of your system’s lifecycle. This is when you want to retire unused data that was migrated as a safety net but never touched in the new system. Tune and refine your reorder points and PAR levels based on actual new system usage patterns — behavior will differ from your legacy system, so fine-tune your processes accordingly. Refine your item classifications and movement classes as you learn how the system processes transactions. And critically, establish recurring data quality audits. This is the governance discipline that prevents the slow drift back toward the data problems you worked so hard to address before go-live. Continue checking in on your governance and data policies to make sure they’re being followed and that people aren’t slipping back into old habits.
A few industry-specific callouts for the post-live optimization phase. For healthcare: focus on inventory accuracy at the clinical level — stockouts and overstock look very different in surgical suites versus general supply rooms. Monitor contract compliance to ensure GPO and local contracts are honored in purchasing and pricing. And work on analytic readiness — building the clinical spend dashboards that were part of the original value proposition. For public sector: focus on procurement compliance, verifying that contract terms and regulatory requirements show up in actual transactions. Make sure your commodity codes are standardized — whether you’re using NIGP or UNSPSC codes, ensuring they’re aligned pays dividends when it comes to spend tracking and transparency. And monitor vendor performance — tracking your highest-spend vendors and how they’re performing against terms.
In both cases, data quality is the foundation you built before and during the project, and it’s what makes all of this achievable.
Let me bring everything back to three principles that underpin every successful data conversion. First: plan early. Organizations that start data readiness work two to three years before a project kicks off have dramatically better outcomes — not because they’re more sophisticated, but because they gave themselves the time to do the work properly. The runway matters. Second: govern well. Governance matters more than tools. I’ve seen organizations with sophisticated platforms fail because no one owned the process. And I’ve seen basic tooling succeed because there was clear ownership, documented rules, and real accountability. People drive success, not software. Third: clean data. At the end of the day, clean data is the foundation of every benefit your ERP investment is supposed to deliver — user trust, operational efficiency, audit readiness, analytics. None of it works without that foundation. Start early, build structure, measure impact. These things compound, and the ROI pays off every day your system runs.
So where do you go from here? Four concrete actions that can move the needle at any phase. First, conduct a data readiness assessment — get a clear-eyed view of where you stand and benchmark against conversion requirements. Second, identify your three to five highest-impact cleansing initiatives — focus on the domains with the greatest return on investment. Third, assign formal data owners — document it, communicate it, hold people accountable. Structure is what makes the difference. Fourth, build data governance into your ERP roadmap from day one — not as an afterthought, but as a foundational milestone. RPI Consultants has done this work across dozens of organizations, and we’d genuinely love to talk about your data strategy. If you have questions or would like to pick our brains, please reach out and let’s have that conversation.
Before we wrap up: if you’re attending the GHX Summit in New Orleans this May, come find us — I’ll personally be there. Whether you want to continue the data strategy conversation, share some war stories, or just grab a cup of coffee, I’d love to connect in person. I’m active on LinkedIn covering supply chain, ERP, and data topics and welcome anyone to connect. Thank you all for being here — I’m happy to take any questions you have.
