ERP Roundup: Advantage Solutions Pauses Its Workday Rollout, and BC29 Admits Your Sandbox Was Never Production
A programme paused rather than rescued, a Business Central action that lends a sandbox a production database configuration for three days, Copilot Business arriving in the price list with a consumption meter behind it, and a runtime deadline four days out. Four stories from the past seven days.
1. Advantage Solutions pauses its Workday implementation, on return rather than defects
On its 4 November third-quarter earnings call, Advantage Solutions CFO Chris Growe said the company had "paused the Workday implementation while we re-evaluated scope, design, and timing against current business needs", and that "the cost of the rollout and anticipated financial and operational benefits no longer delivered an adequate return". The existing platform stays in use in the interim, and Growe added that "importantly, this decision reduces our 2026 capital spending expectations".
Two details are worth keeping straight. The call describes a Workday implementation and an existing "human capital platform"; the transcript publisher's summary calls it an ERP system, and no module scope was named, so how much of this was finance rather than HR is not on the record. And nothing said alleges that the software failed. This is a buyer withdrawing a business case, not a go-live that broke.
What it means for testing
A programme that reaches "no longer delivered an adequate return" has usually been re-scoped several times first, and every re-scope moves the acceptance criteria, informally, in a workshop nobody minuted. The defence is to hold acceptance scope as a tracked artefact, so that when the question becomes whether the remaining scope justifies the remaining spend, it can be answered from a record: the same discipline as treating sign-off as a dated decision rather than an email, applied earlier.
2. BC29 lends your sandbox a production database configuration — for 72 hours
Yun Zhu documented on 4 November a Match Production Database Configuration action that has appeared in the Business Central admin centre. It temporarily gives a sandbox the database configuration of a typical production database for 72 hours, so that Performance Toolkit results mean something; the environment restarts, and reverts in the next update window. It appears in neither the BC29 release plan nor the AL Language v18.0 changelog, and Zhu's own attempt in a paid environment returned "Production database configuration availability couldn't be determined". He concludes it "appears not yet ready for practical use".
The feature matters less than what shipping it concedes. Production environments run on dedicated performance tiers and sandboxes do not, so every performance figure anyone has taken in a Business Central sandbox was measured on different hardware from the one the business will use. A button in the admin centre makes that a documented property of the platform — though for now, like the Fabric mirroring preview, it is described ahead of being usable.
What it means for testing
If UAT runs in a sandbox, your throughput numbers are not production numbers and should not go into an exit criterion as though they were: stop accepting performance criteria on sandbox evidence, or record on the sign-off that the figure is indicative. And a 72-hour window is a weekend, not a test phase — scripts, data volumes and a named operator must exist before it opens, which is ordinary UAT preparation discipline applied to a clock you do not control.
3. Copilot Business enters the Business Central price list, with a consumption meter behind it
A Partner Center announcement dated 5 November introduces two permanent CSP bundles pairing Business Central Essentials or Premium with Microsoft 365 Copilot Business, for customers up to 300 seats. Essentials with Copilot Business is USD 95 per user per month against USD 101 bought separately; Premium is USD 125 against USD 131. They hit the price list preview on 1 December 2026 and sell from 1 January 2027, with no qualifying Microsoft 365 base subscription required — though Microsoft's caveat is that "Copilot experiences continue to depend on the Microsoft 365 apps and services a customer is licensed for". Note this is Microsoft 365 Copilot Business, not Business Central's in-product Copilot.
Two days earlier, the same workspace moved the usage-based billing default for new Copilot Business purchases in CSP from 1 December 2026 to Microsoft's FY27 Q3, January to March 2027. The default spending limit is 4,000 Copilot Credits per user per month, adjustable by admins; Microsoft's worked example is 100 users consuming up to 400,000 credits a month, billed on actual usage. Disabling it, the page says, "might limit access to AI experiences".
What it means for testing
A per-user credit ceiling is an acceptance criterion in a shape most UAT scripts have never carried. The question stops being whether the feature returns the right answer and becomes what one run consumes, multiplied by the volume the business actually does — measurable during UAT, almost never measured, because nobody owns it. The dates are decoupled too: the bundle sells from 1 January, the billing default changes in the quarter after. The cycle that absorbs that guesswork is the one with part-time stakeholders.
4. .NET 8 support ends for Marketplace SaaS accelerators on 10 November
Also dated 5 November, flagged action required for "partners with active SaaS Marketplace offers": the latest update to the open-source SaaS accelerator for Microsoft Marketplace adds .NET 10 support and ends support for .NET 8 on 10 November 2026. Update "as soon as possible to avoid disruption", Microsoft says, because "failing to update could result in failed or delayed Marketplace transactions" — and partners running several accelerator instances across offers need to update all of them.
The accelerator handles Marketplace subscription and billing webhooks for transactable offers, not the product itself. So for an ISV with a Business Central app on Marketplace, the exposure is the commerce path: the money, not the features.
What it means for testing
Four days to move a runtime under the component that takes payment. "Does it still start" is not the test; the subscription lifecycle is — purchase, activation webhook, plan change, suspend, reinstate, cancel — and those are the paths nobody re-runs, because each fires once per customer and never in a demo. Monday's post ran this arithmetic for SAP security notes: the fix takes a day, proving it did not break order-to-cash takes weeks. Here it is four days, and the untested path is the one that invoices.
The through-line: the environment is part of the test
Three of this week's four stories change the conditions a test runs under rather than the behaviour being tested: the database configuration the sandbox sits on, the credit ceiling a feature consumes against, the runtime the commerce path transacts on. None is a defect, and none would appear on a test script. The fourth is why that matters: when a programme's numbers stop working, what gets withdrawn is the benefit case, and the acceptance record is the only durable account of what it was meant to prove. Last week it was a quarter's revenue; this week a capital budget.
Free ERP UAT checklist (Excel)
A structured workbook covering the full UAT cycle: pre-UAT preparation tasks, test execution tracking, an issue log with severity guide, and post-UAT wrap-up. Built from real ERP implementations.
No spam. We will occasionally send UAT and ERP implementation resources. Unsubscribe any time.
Sources and further reading
- Advantage Solutions (ADV) Q3 2026 earnings call transcript, in which CFO Chris Growe states that the Workday implementation has been paused pending re-evaluation of scope, design and timing, that the rollout cost and anticipated benefits no longer delivered an adequate return, and that the decision reduces 2026 capital spending expectations (4 November 2026)
- Yun Zhu (Microsoft MVP), Dynamics 365 Lab: Match Production Database Configuration in Business Central 2026 release wave 2 (BC29), a 72-hour temporary match of a sandbox to a production database configuration, found in the admin centre and the BC Launch Event video but absent from the BC29 release plan and the AL Language v18.0 changelog (4 November 2026)
- Microsoft Partner Center: November 2026 announcements, covering the two Business Central plus Microsoft 365 Copilot Business CSP bundles at USD 95 and USD 125 per user per month for up to 300 seats, with price list preview on 1 December 2026 and availability on 1 January 2027 (5 November 2026); the usage-based billing default move to FY27 Q3 with a 4,000 Copilot Credits per user per month default spending limit (3 November 2026); and the end of .NET 8 support for Marketplace SaaS accelerators on 10 November 2026 (5 November 2026)
When The Business Case Changes, The Acceptance Record Is What Survives
LogicHive keeps the acceptance layer where a re-scope cannot quietly erase it: structured test cases tied to requirements, named business owners, linked defect tracking, and sign-off built from real execution data rather than a screenshot of a passing build. When a programme is paused, a performance figure is questioned or a consumption bill arrives, the record of who accepted what, on what evidence, is already there. The free plan runs a real project end to end.
Written by Sharif George
LogicHive Frontman
Sharif is the frontman for LogicHive. He writes the ERP news roundups and the Business Central release coverage, following each Microsoft, SAP and Oracle change through to what it means for regression scope and acceptance testing on live implementation projects.
View all articles by Sharif GeorgeRelated reading
ERP & Enterprise
SAP Patches Get Exploited in Days Now. Regression Testing Still Takes Weeks.
Onapsis Research Labs and Recorded Future published Days to Zero on 29 October: for SAP vulnerabilities patched in 2025 and 2026, the median gap between a Security Note being released and verified public exploitation or a published proof of concept was one to three days. For 2021 patches it was 489 days. The caveats matter — the median combines exploitation with proof-of-concept publication, the stated range since April 2025 is 0.3 to 24 days, and no sample sizes are published. The direction is not in doubt, and it turns the patch decision into a testing-capacity decision: can you prove, in days, that the patch did not break order-to-cash?
2 November 2026 · 5 min read
ERP & Enterprise
Microsoft Moved Sustainability Into the ERP Portfolio. No UAT Plan Has a Tester Who Can Sign Off a Carbon Number.
On 3 November, Microsoft's Chief Product Officer for Dynamics 365 Finance, Human Resources and Sustainability wrote that “we are welcoming Microsoft Sustainability Manager into the Dynamics 365 ERP portfolio”. No feature in the post carries a preview or general availability date, and the green ledger view it describes was already listed in the 2026 wave 1 release plan — so the announcement consolidates ownership rather than shipping anything. What it does change is scope. Emissions figures are produced by configured calculation models: conditional logic, estimation factors, emission factor libraries and custom Power Fx with documented gaps. They fail by returning a plausible number rather than an error, and on a typical ERP project nobody is nominated to approve them.
9 November 2026 · 5 min read
ERP & Enterprise
ERP Roundup: LKQ Puts $95 Million of Lost Revenue on an ERP Disruption, and BC29 Deletes a Page Microsoft's Docs Still Describe
LKQ's CFO told the 29 October earnings call that revenue loss in Europe due to the ERP was approximately $95 million and that the disruption in Germany cut EBITDA by about $43 million in the quarter — six days after the company withdrew guidance and lost its chief executive. Alongside it: the Custom Report Layouts page is gone in BC29 while Microsoft's own documentation still calls the feature merely obsolete, AL interfaces gain default method bodies your build can inherit silently, and the SMB referral acceptance window drops to five calendar days on 2 November. Four stories, and in every one the cost lands a long way from the decision that caused it.
30 October 2026 · 7 min read