B2B Risks and Dependencies
Shareable with BLACKROLL and project partners. Maintained by 22Bits. Written 2026-10-16, updated after the kickoff 2026-10-21, after a call with Sen 2026-10-22, after emails from Sen and Kristy 2026-10-23 and after the B2B weekly 2026-10-26. Part of Project: B2B Migration.
What the project depends on that is outside the B2B store or outside 22Bits, what everyone seems to assume without having written it down, and the main risks for the January date. Each item points to the decision in B2B Decisions and Workshop Plan that addresses it.
Dependencies and unwritten assumptions
ID | What seems to be assumed | What the sources say | Status | Decision |
|---|---|---|---|---|
D1 | That Xaidat can scope and deliver Option B in time | Jira BLAC-6174 and BLAC-6176 read like Option B ("One source of truth for prices, synced from ERP", "Integrate through the Xaidat gateway"), and BLACKROLL prefers it. Xaidat is a third party and has not given a scope, estimate or timeline yet. Option A stays the fallback for the CH wave | Open | E1 |
D2 | That CembraPay will install on the new store once the currency is chosen | Open since ADR-060 (July); the Finance clarification was assigned at the workshop of 2026-07-16 and no result is recorded. At the kickoff a workaround was mentioned and a breakout agreed. Sen (2026-10-23) prefers replacing CembraPay with native payment terms, if the Swiss requirements are covered | Open | PC1, PC3 |
D3 | That the three legacy shops are known | Resolved at the kickoff: CH store, Pro DE store, Pro BNL store. The Pro BNL store is an English-language store with daily orders, so BNL is a migration, not a market launch. Sen (2026-10-23) also names France, which is in no wave yet (question 25) | Resolved, France open | S3, S8 |
D4 | That the legacy CMS is a separate CMS | It is the old Craft deployment, read only by the legacy Nuxt on Vercel, which now serves nothing but the CH B2B fallback (all D2C markets are on the new instance and the new CMS). Its switch-off follows the CH wave; only the D2C cleanup children sit under BLAC-6012 | Open | M3 |
D5 | That Brand 2.0 components can be "reused" in Liquid (BLAC-6179) | They are Vue, Tailwind and Storybook; a Horizon theme gets tokens and layouts re-implemented, and each reused component is a rebuild. No designer is assigned (confirmed by Sen, 2026-10-23) | Assumed: tokens and typography only, needs clearance | T2 |
D6 | That dealer data is clean enough to map | The spike measured 31% multi-address customers, 33 restriction combinations and untiered customers paying list price; according to ADR-060 CH dealers have no ERP record at all, while Sen (2026-10-23) describes the CH process as Shopify first, then MyFactory (question 26); owners and meaning of many customer tags and metafields are unknown | Open | CI3, CI4, P4, E3 |
D7 | That the DEPT handover has a receiver | BLACKROLL's operating model has DEPT run the platform from February; the DEPT Shopify expert who wrote the spike is marked non-active; the runbooks in BLAC-6189 describe D2C infrastructure. Accepted on 2026-10-26: no name in November; DEPT assigns the receiver in December, and the handover goes to whoever is assigned. That person will have been in none of the workshops, so the handover plan, the runbooks and this wiki have to carry the context | Open | PL5, S5 |
D8 | That b2b.blackroll.com is free | It serves the legacy CH shop since 2026-09-18 and has to keep doing so until the CH wave is live. At the kickoff Sen suggested keeping pro.blackroll.com | Open | PL2 |
D9 | That the CH bridge on the new D2C store is in place | BLAC-5366 (map the B2B flag during migration) is still blocked; per Kristy (2026-09-17) the B2B customers were not migrated to the new platform. The fallback (ADR-060 scenario 3) is what went live, not the bridge (scenario 1). The deck still calls it the "voucher bridge" | Open | M1, question 10 |
D10 | That the Pros question can wait | It changes the wave 1 cohort and the data migration mapping (23,486 CH segment members versus 700 to 1,000 companies; about 6,800 B2B partners in the Pro shop). Kristy asked at the kickoff to front-load it | Assumed: Pros into the B2B store | S2 |
D11 | That a Liquid store needs Cloudflare and Datadog (slide 6, BLAC-6189) | Shopify advises against a proxy in front of the storefront; the runbooks would describe things the store does not have | Open | PL5 |
D12 | That 22Bits is in WS7 (CRM refactor) | The deck says Philipp, Marc and Front Row; the epics say Front Row (6184, 6188) and 22Bits plus Front Row (6183); the integration platform choice (Make or n8n) is under review, and Klaviyo itself was questioned at the kickoff | Open | PL6, CI6 |
D13 | That the price model stays as it is while Xaidat builds Option B | A new B2B price model ("B2B Direct Price") for the B2B shops and EDI is being worked out (Sen, 2026-10-21). Sen aligns it internally with Sebastian Krebs before workshop 3b (2026-10-23). Since 2026-10-26 prices live only in Shopify (P2), so Xaidat no longer builds price lists and the risk of doing that work twice is gone. The new model is built in Shopify catalogs; the price sync moves to the integration with the new ERP later (D17) | Mostly resolved by P2 | P1, P2, E1 |
D14 | That a template-based theme meets the business expectation for the look | The B2B team would like the store closer to the D2C look, with D2C content (Sen, 2026-10-21). The working assumption is tokens and typography on Horizon's B2B blocks; each reused D2C component is a rebuild in Liquid | Open, direction in workshop 0 | T1, T2 |
D15 | That Sen is available to decide and confirm throughout the build | Sen is out of office from 13 to 23 November and from 21 December to 2 January (Sen, 2026-10-23). The first absence falls on store setup and the CH companies and price lists, the second on the CH dealer communication. Without a deputy, decisions from workshops 1 to 3 that are not confirmed by 12 November wait until 24 November | Open | S5, M2 |
D16 | That product data for the B2B store stays as it is during the CH wave | A new PIM is planned for January (kickoff), the same month as the CH acceptance test. Sen wants the D2C product data as the base for B2B, plus B2B-specific fields (2026-10-23). If the PIM switch changes the product feed, the B2B store has to follow it | Open | E6, E7 |
D17 | That the new ERP comes after this project and can take over from the B2B store | Kristy (2026-10-26): a new ERP comes later, and price management moves into that integration. Which ERP and when is not known (question 27). The B2B store keeps prices, catalogs and companies in standard Shopify objects, so a later integration can write to them through the Admin API | Open | P2, E1, E4 |
Risks for the January date
Written as a pre-mortem: it is March 2027 and the project is in trouble. What went wrong?
ID | Risk | How it happens | What prevents it |
|---|---|---|---|
R1 | The CH wave slips past the next sale window | Store currency and CembraPay are decided late, and every price list has to be re-entered after a currency change (Shopify changes the currency label, not the amounts, store currency). The store type is the other one-way door: "If you set up one type of store and then want to switch to the other type, then you'll need to redo most of your setup work" (store type) | PC1, PC3 and PL1 decided before the stores are created; workshop 2 in the week of 2 November |
R2 | The scope is never signed | All 20 epics stay "in", priorities change every week, and January ends with a half-built DE wave and a CH wave in hypercare. The estimate gate on slide 9 is never passed | S1 with the wave approach in the first week of November; a provisional estimate on the working assumptions, updated after each workshop |
R3 | Custom code without an owner after January | The tag translator app or the migration scripts have no one at DEPT who has seen them before December, and the first incident (a dealer with a missing price list) waits a week | PL4 (as little custom code as possible), PL5 (the receiver DEPT assigns in December gets the handover plan), runbooks written while building, and a handover plan written for someone who was in none of the workshops |
R4 | Dealers are told late or wrong | The comms plan per wave depends on a sales lead who is not named yet; the dealer application form lives in two systems with different approval paths | M2, CI1, S5 |
R5 | The data migration reconciles on paper, not in practice | Duplicates across the CH segment and the DE ERP list, multi-address companies modelled as single locations, restricted customers without their catalogs | CI3, CI4, P4, dry runs before each wave |
R6 | The theme scope grows | Brand 2.0 turns into a rebuild of D2C components and design review cycles on a Liquid theme without a designer, while Horizon's B2B blocks would have done | The T2 direction decided in workshop 0, then T3 and question 17 |
R7 | The legacy CMS can't be switched off | The CH wave slips and the legacy Nuxt on Vercel keeps serving the B2B fallback, and the Cloudflare router's legacy sitemap fallback (BLAC-3340) is never removed | M3, question 18 |
R8 | Content ownership falls between teams | Nobody owns the question of which content is shared with blackroll.com, and dealer content ends up typed twice | T1 and its clearance; content ownership named in BLAC-6181 |
R9 | Custom development eats the budget | Every must-have that standard Shopify does not cover gets its own app or script; together they cost more than the budget allows and leave DEPT more to run | The guiding principle (Kristy, 2026-10-26): out-of-the-box Shopify first; PL4; Sen's wave marking of the requirements list (S1); a cost note on every lean in the decisions page |