Every Dynamics 365 CE (D365 CE) practice has a version of the same story.
The proposal goes out. Scope is defined. The data migration is in there: extract, transform, load, validate. Duplicate detection is in there too, at least implicitly. The client assumes clean data on day one. The practice assumes the standard workaround will hold.
Go-live week arrives. The workaround takes longer than estimated. A plugin needs adjusting for a custom entity nobody scoped. The bulk cleanup job hits the 5,000-record ceiling and has to be batched manually. The consultant who built the original workaround is now splitting time between this project and the next one.
The project closes. The client signs off. The duplicate workstream absorbed six hours more than estimated. On a fixed-fee engagement, those six hours came out of margin.
Multiply that across a year of migrations, and the duplicate workstream stops being a project inconvenience. It starts being a line item that nobody is tracking.
A boutique CE practice running five to eight migrations a year isn't unusual: mid-market clients, Dynamics 365 CE implementations, a full delivery cycle that includes data migration. Each project is scoped at a fixed fee and delivered by a team of two or three consultants.
On each of those projects, the duplicate workstream involves some combination of:
Conservatively, that's eight to twelve hours per project when the workaround behaves as expected. When it doesn't, a new integration connector, a custom entity that wasn't in the plugin scope, a larger-than-expected dataset, the hours climb.
At eight projects a year, the low end is 64 hours. The high end, on projects where something unexpected surfaces, is closer to 120.
The 64-to-120-hour range only captures the work that gets logged. There's a second category that doesn't.
Tribal knowledge transfer. The consultant who built the workaround on project one has context that isn't documented anywhere. When a new consultant joins project four, someone has to get them up to speed. That handoff happens informally, over Slack or a quick call, and it doesn't go on a timesheet. It comes out of someone's afternoon.
Go-live firefighting. When duplicates surface at go-live, the response is rarely calm and methodical. The consultant who owns the data workstream gets on a call with the client, pulls up the environment, and diagnoses in real time. That hour or two isn't pre-scoped. It's absorbed.
Post go-live cleanup requests. Clients find duplicates after sign-off. They raise a support ticket or call the practice lead directly. The practice decides whether to treat it as a warranty item or bill for it. Either way, someone does the work.
Add those categories together, and the annual cost of the dedupe workaround starts to look less like a rounding error and more like a mid-level consultant's monthly capacity.
Time-and-materials engagements can absorb overruns. Fixed-fee engagements can't, not without a conversation about scope.
Most boutique CE practices run fixed fees. It's what clients want, and it's what separates a practice that knows its delivery well from one that doesn't. The risk of a fixed-fee model is that every unscoped hour comes directly off the bottom line.
The duplicate workstream is almost always underscoped, not because practice leads are careless, but because the workaround behaves differently on every project. The entity model is different. The integration connectors are different. The dataset size is different. The same workaround that took four hours on the last project can take ten on this one, and there's no reliable way to predict which it'll be at proposal time.
The result is a cost that gets absorbed rather than billed, repeated across every engagement, and never quite large enough on any single project to force a conversation about it.
Using a conservative billing rate of $100 an hour, low for a senior consultant at a boutique practice, and a mid-range estimate of ten unscoped hours per project:
These aren't dramatic numbers in isolation. But they represent margin that was earned and then given back, on work that doesn't differentiate the practice, doesn't build client relationships, and doesn't make the next project any easier.
The number also compounds. A practice running eight migrations a year for three years has absorbed somewhere between $20,000 and $30,000 in duplicate workstream overruns, conservatively, without a single internal conversation about whether there's a better way to handle it.
The hours absorbed by the dedupe workaround have a second cost that doesn't show up in any margin analysis: what those hours could have been used for instead.
A senior consultant spending ten hours on duplicate cleanup is ten hours not spent on configuration work, business process design, or client-facing delivery that the client notices and remembers. It's ten hours not spent on the next project, on internal practice development, or on the kind of work that generates referrals.
The dedupe workaround is infrastructure work. Infrastructure should be invisible: configured once, running in the background, not requiring a consultant's attention on every engagement. When infrastructure work consumes senior consulting hours on every project, the practice is subsidizing a problem that already has a solved alternative.
The practices that have moved away from the per-project workaround describe the same shift: the duplicate workstream goes from a variable cost to a fixed, predictable one.
A native Dataverse deduplication layer, installed from Microsoft Marketplace, gets configured once per client environment and covers every entry point without per-integration setup. Standard and custom tables are both included. There's no 5,000-record ceiling on bulk jobs, and the configuration stays consistent across engagements instead of being rebuilt from scratch each time.
Plauti works this way inside Dynamics 365 CE: it installs inside the client's Dataverse environment, runs without external dependencies, and catches duplicates at every entry point from day one of go-live. For a practice running five or more migrations a year, the licensing cost is typically recovered within the first one or two projects where the workaround would otherwise have overrun.
The margin doesn't disappear. It stays where it was earned.
Before the next migration gets scoped, it's worth asking one question internally: how many hours did we absorb on the duplicate workstream on the last three projects, and what did we estimate?
If the answer is "we don't track that separately," that's the first thing worth changing. The cost is there. It just isn't visible yet.
Blog
Most Dynamics 365 CE partners solve duplicate detection the same way on every migration, then rebuild it from scratch next time. Here's why, and what a one-time platform fix looks like.
Blog
Dynamics 365's native duplicate detection only catches UI form entries. Bulk imports, Web API calls, and Dual-Write syncs bypass it by default, which is why duplicates pile up right at go-live.
The duplicate workstream behaves differently on every project: different entity models, different integration connectors, different dataset sizes. A workaround that took four hours on the last project can take ten on the next one, with no reliable way to predict the variance at proposal time. On fixed-fee engagements, that variance comes directly out of margin.
A conservative estimate for a well-behaved workaround is eight to twelve hours per project, covering plugin configuration, integration header setup, post-migration bulk detection batching, and initial cleanup. On projects where something unexpected surfaces, such as a new connector, an out-of-scope custom entity, or a larger dataset, the hours climb significantly. Go-live firefighting and post-sign-off cleanup requests add further unlogged time on top.
For a practice running eight migrations a year at a conservative billing rate of $100 an hour, ten unscoped duplicate workstream hours per project amounts to $8,000 absorbed annually. That figure doesn't include informal knowledge transfer, go-live firefighting, or post-sign-off cleanup, all of which add unlogged hours. Practices running ten or more migrations a year can absorb $10,000 or more annually in duplicate workstream overruns.
Yes. Replacing the per-project workaround with a native Dataverse deduplication layer converts a variable, unpredictable cost into a fixed licensing cost. A tool installed once per client environment, covering all entry points and all entity types without per-integration configuration, removes the variance that makes the workaround hard to scope. For practices running five or more migrations a year, the licensing cost is typically recovered within the first one or two projects where the workaround would otherwise have overrun.
With a product tour you can walk through the product yourself without installing anything, or book a demo for a guided look.