Deeptech founders' guide to multi-SKU hardware costing and portfolio cash flow modelling
Why Multi-SKU Hardware Portfolios Complicate Costing
The first hardware product is a success, and the roadmap is clear: it is time to expand the portfolio. A second SKU, then a third, is planned to capture more of the market. Yet the financial clarity that came with a single product begins to fade. The straightforward calculation of unit cost becomes complex, and the cash flow forecasts that once seemed reliable now feel unpredictable. This isn't a sign of failure; it is a predictable growing pain for hardware startups scaling their product lines. Understanding how to track hardware development costs for multiple products is essential for survival. This guide provides a framework for Deeptech founders to regain financial control. For broader context, see the Hardware NPI Costing & Capex topic.
Diagnosing the Fuzziness: How Unit Economics Get Blurry with Every New SKU
For a single product, calculating profitability is relatively direct. You have your non-recurring engineering (NRE) costs, your capital expenditures (Capex) for tooling, and your per-unit Bill of Materials (BOM). When a second product variant arrives, especially one that shares a core platform, this simplicity evaporates. The primary challenge is distinguishing between shared platform costs and SKU-specific costs.
A common but dangerous approach is to average, or “peanut butter,” all development and tooling costs across all units sold, regardless of the SKU. This method is tempting because it is simple, but it obscures the truth and leads to poor strategic decisions. It blurs the true profitability of each product, potentially causing you to overinvest in a low-margin product or prematurely kill a future winner. The reality for most Deeptech startups is more pragmatic: you need to see which products carry their own weight and justify their existence financially.
Consider a smart home company developing two products from a shared platform:
- Core Platform: A shared WiFi module with a research and development cost of $200k.
- SKU A: A smart thermostat requiring a unique plastic housing mold that costs $50k.
- SKU B: A smoke detector with its own housing mold costing $40k and a separate, mandatory UL certification costing $30k.
You can learn more about planning large mold expenses in our tooling investment planning: ROI framework.
If the company lumps all $320k of NRE and Capex together, it misrepresents each product's financial burden. SKU B has $70k in specific costs, while SKU A only has $50k. Averaging these hides the fact that SKU B requires a higher gross margin or greater sales volume to justify its unique certification and tooling investment. The first step in any effective hardware product cost breakdown is to isolate what is truly shared from what is unique. A clear understanding here is fundamental to effective hardware development budgeting and detailed product variant cost analysis.
The Capex Blind Spot: How to Track Hardware Development Costs for Multiple Products to Avoid Cash Crunches
Even with a clear budget for each product, hardware startups often face unexpected cash shortages during production ramp-ups. This happens because per-product budgets obscure the cumulative cash outflow required to launch multiple products in parallel or in quick succession. A budget shows what you will spend; a cash flow forecast shows when you will spend it.
A scenario we repeatedly see is the convergence of large, one-off payments. Imagine a situation where payments for three $50k molds plus two sets of test equipment hitting a bank account in a 60-day window. A per-product view shows three distinct, manageable expenses. A portfolio-level cash flow forecast, however, reveals a massive, concentrated cash demand of over $200k that could jeopardize runway. This is the critical distinction between per-product budgeting and cumulative cash flow modeling. Your budget might be sound, but your timing could be fatal.
This problem is amplified by the “forgotten Capex” of launch inventory. The first large production run is often the single biggest cash outlay for a hardware company. While technically accounted for as Cost of Goods Sold (COGS) when the units are sold, the cash leaves the building months before any revenue arrives. For companies in the UK or USA, this cash is gone, and you must survive until sales replenish it. When managing multiple SKU expenses, this means forecasting the stacked inventory payments for all new products simultaneously.
Without this portfolio-level cash flow view, even well-funded startups can find themselves unable to pay suppliers, stalling production and delaying revenue. This is a classic challenge in capex planning for hardware startups. The challenge is often made worse by supplier terms, such as large minimum order quantities (MOQs) for critical components, which force even larger upfront cash commitments. Recent R&D tax relief reforms in the UK may also affect the timing and amount of cash available from tax credits.
The Solution: Building a Dynamic Model for Hardware Portfolio Financial Management
To regain financial control, founders need a simple, dynamic tool for hardware portfolio financial management. This doesn't require an expensive ERP system or a full-time finance hire. At the Pre-Seed to Series A stage, a well-structured spreadsheet is perfectly sufficient. The goal is to move from reactive accounting in a tool like QuickBooks or Xero to a proactive financial strategy.
The model is built on three distinct cost buckets:
- Shared NRE & Capex: This bucket contains all one-time costs that benefit the entire product portfolio. This includes foundational R&D for a core technology, shared testing equipment (like an RF analyzer), and platform-level regulatory certifications like FCC or CE that cover the shared modules.
- SKU-Specific NRE & Capex: These are one-time costs tied directly to a single product. Examples include a unique mold for one product’s housing, like the SKU A (Thermostat) unique plastic housing mold cost example: $50k, or a product-specific certification, such as the SKU B separate UL certification cost example: $30k.
- Variable Unit Costs (COGS): This bucket includes all per-unit costs. It starts with the Bill of Materials (BOM) and also includes assembly labor, packaging, logistics, yield loss allowance, and warranty accrual.
Founders should be aware that accounting standards such as IAS 38 (and its US GAAP equivalent) provide specific rules on which development costs can be capitalized versus expensed.
With costs categorized, the next step is to strategically allocate the Shared NRE & Capex bucket across the SKUs. A simple and defensible method is allocation based on forecasted lifetime unit volume. This approach assigns costs based on the expected contribution of each SKU to the business over its life, avoiding the trap of penalizing a new product with high upfront costs before it has a chance to gain market traction.
Let’s apply this to our smart home example. The core WiFi module shared R&D cost example: $200k. The company forecasts selling 100,000 total units over the products’ lifetimes, with an example allocation based on forecasted lifetime unit volume: 70k units of SKU A (70%) and 30k units of SKU B (30%). The allocation is straightforward:
- SKU A (Thermostat) is allocated 70% of the shared R&D: $140k.
- SKU B (Smoke Detector) is allocated 30%: $60k.
Now, a true profitability picture emerges:
- SKU A Total NRE/Capex: $140k (shared) + $50k (specific) = $190k
- SKU B Total NRE/Capex: $60k (shared) + $40k (mold) + $30k (cert) = $130k
The model’s power comes from its ability to answer “what if?” questions. In your spreadsheet, you can create input cells for key assumptions. For instance, you can model a scenario where the main chipset price increases by 20%. By changing that single input, you can instantly see the impact on the gross margin for every SKU and on your total company runway. This dynamic analysis is invaluable for setting prices, negotiating with suppliers, and having data-driven conversations with investors.
Key Actions for Implementing Your Portfolio Cost Model
Moving from a single product to a multi-SKU portfolio introduces financial complexity that can easily overwhelm an early-stage team. Gaining control doesn't require enterprise software, just a disciplined approach. What founders find actually works is building a simple but powerful model that provides clarity on profitability and cash flow. To implement this, focus on four key actions:
- Separate Your Costs Diligently: In a spreadsheet, create the three distinct buckets: Shared NRE/Capex, SKU-Specific NRE/Capex, and Variable Unit Costs. This first step alone will bring significant clarity to your hardware development budgeting.
- Allocate Shared Costs Strategically: Avoid the temptation to average shared costs equally. Use a defensible method, like forecasted lifetime volume, to assign shared NRE to each SKU. This gives you a true understanding of each product’s financial burden and profitability potential.
- Model Cash Flow, Not Just Budgets: Your most important financial tool is a timeline of cash payments. Map out when large Capex invoices for molds, test equipment, certifications, and initial inventory are due. This reveals the crunch, not the budget. It is how you truly manage runway. Our guide on Hardware Working Capital Planning: Inventory Cycles can help you model these stacked inventory payments.
- Build a Simple Scenario Planner: Your cost model should be dynamic, not static. Link your key assumptions, like component costs, sales volumes, and foreign exchange rates, to a dedicated set of input cells. This allows you to quickly see how market shifts affect your gross margin and cash position, turning your model into a strategic decision-making tool.
Your accounting software, whether QuickBooks for US companies or Xero in the UK, is excellent for tracking what has already happened. This forward-looking portfolio model is what allows you to shape what happens next. To learn more, explore the full hub on Hardware NPI Costing & Capex.
Frequently Asked Questions
Q: How often should we update our multi-SKU cost model?
A: Your model should be a living document. We recommend a full review on a quarterly basis and a quick update before any major strategic decision, such as signing a large manufacturing order, changing pricing, or entering investor conversations. This ensures your decisions are always based on the most current data.
Q: What is the most common mistake in allocating shared hardware development costs?
A: The most common mistake is allocating shared costs based on current-period sales volume or revenue. This unfairly punishes new products that have not yet scaled, making them appear unprofitable and potentially leading to their premature cancellation. Allocation based on forecasted lifetime volume provides a much fairer and more strategic view.
Q: How should we account for ongoing software R&D in this model?
A: Treat software R&D like hardware R&D. Core platform development, such as a shared connectivity stack or cloud backend, should be placed in the "Shared NRE" bucket. R&D for features exclusive to a single SKU, like a unique algorithm for a specific sensor, should be categorized as "SKU-Specific NRE."
Q: When is it time to graduate from a spreadsheet to a dedicated ERP system?
A: Transition from a spreadsheet when its limitations create significant operational risk or inefficiency. This tipping point typically occurs post-Series B, or when your business involves multiple contract manufacturers, complex international supply chains, and a rapidly growing number of SKUs that make manual tracking unreliable.
Curious How We Support Startups Like Yours?


