BOM cost reduction in electronics means lowering what you pay for the parts on a bill of materials without changing what the product does. There are eight levers, and they differ mainly in who has to do the work. Four are purchasing levers: price breaks, packaging, landed cost and timing. They cost no engineering time and pay back on the next order. The other four need an engineer to sign off: value consolidation, grade right-sizing, a pin-compatible second source and a part downsize. Those only pay back above a volume you can calculate before anyone opens a schematic. As of October 2026 the order of work matters more than usual, because memory, analog and passive prices are still rising into the fourth quarter.
Which BOM lines should you attack first?
Start with the lines that carry the most spend, not the most parts. To find them, multiply each line's unit price by its annual volume and sort descending. On a typical industrial controller board, the processor, the memory, the power stage and the connectors sit at the top of that list. A 10% saving on any of them is worth more than halving the price of fifty resistors. A line costing EUR 6.40 at 8,000 units a year is EUR 51,200 of spend. A EUR 0.004 resistor at 40 placements per board is EUR 1,280.
Then flag the lines that face a price change you already know about. TrendForce projects conventional DRAM contract prices to rise 10-15% quarter on quarter in Q4 2026, and NAND Flash 15-20% (TrendForce, 30 September 2026). Analog and passive suppliers have published their own increases, which we track in our 2026 price increase chart. On a line with a dated increase, the cheapest saving is often buying before the effective date. Redesigning it comes later, if at all.
The eight levers, ranked by payback
Purchasing levers come first because they pay back on the first purchase order. Engineering levers follow in order of rising requalification effort. The table ranks them by how quickly the saving covers the work, not by how large the saving can get. A part downsize can save the most per unit and still rank last.
| # | Lever | Who does the work | Requalification | Pays back when |
|---|---|---|---|---|
| 1 | Order to the next price break or full reel | Purchasing | None | The surplus is used within its shelf life |
| 2 | Right packaging format (reel, cut tape, tray) | Purchasing | None, if the base part number is unchanged | Volume crosses one full reel |
| 3 | Landed-cost check (duty, freight, MOQ waste) | Purchasing | None | Immediately, on any cross-border order |
| 4 | Buy ahead of a dated price increase | Purchasing and finance | None | Increase x volume exceeds holding cost |
| 5 | Consolidate duplicate passive values | Engineering, light | Design review only | Fewer lines reach full-reel quantities |
| 6 | Right-size the grade (automotive, temperature) | Engineering and quality | Spec check, no new testing if electrically identical | The product never ships into an AEC-Q100 contract |
| 7 | Pin-compatible second source | Engineering | Bench test and first-article | Annual saving beats roughly 40 engineering hours |
| 8 | Downsize to a cheaper part in the same family | Engineering | Firmware, layout check, full validation | Volumes in the tens of thousands per year |
Purchasing levers: what to change on the next order
Each of the four purchasing levers changes how you buy, not what you buy, so none needs engineering approval. Together they are the fastest saving on most BOMs.
Price breaks. Price breaks do not stack. Buying 2,000 pieces in four lots of 500 pays the 500-piece price four times. Before a release, check whether the next break costs less in total than the quantity you need. For a part needed at 2,300 pieces with a 2,500-piece break, it often does. Buy the surplus only if you will use it within the part's storage life. Moisture-sensitive packages in an opened bag have the shortest limits.
Packaging. Cut tape carries handling cost, and some distributors add a reeling fee for custom reels. Once annual volume passes one full reel, the reel suffix is usually the cheaper code for the same die. The tape and reel vs cut tape guide has the suffix map and a worked comparison.
Landed cost. Unit price is not what a European buyer pays. As of October 2026 the EU applies 0% third-country duty to integrated circuits (HS 8542), transistors and diodes, capacitors and fixed resistors, but connectors and relays (HS 8536) carry 2.3% and cable assemblies up to 3.7%. A connector quote that is 2% cheaper from a supplier outside the EU loses that margin to duty alone, before freight and any minimum order quantity you do not need. Our HS code guide lists the rates by component family.
Timing. When a supplier has published an increase with an effective date, compare the increase on your forecast volume with the cost of holding that stock. A 12% DRAM increase on EUR 40,000 of annual spend is EUR 4,800. Buying the year's volume ahead and holding it for six months on average, at a 10% annual carrying rate, costs about EUR 2,000. The early buy saves about EUR 2,800 net.
Engineering levers: when a part change is worth the work
An engineering change pays when the annual saving exceeds the one-time cost of qualifying it. The quickest way to see this is to put a number on that cost before arguing about the saving. Our working assumption for a pin-compatible swap on a non-safety industrial board is 40 engineering hours at EUR 95 per hour, plus EUR 1,200 for samples, bench testing and documentation. That comes to about EUR 5,000 per swap.
- A EUR 0.12 saving at 8,000 units a year returns EUR 960 a year, so the swap needs five years to pay back. Leave it.
- A EUR 0.85 saving on a power stage at the same volume returns EUR 6,800 a year and pays back in nine months. Do it.
- A downsize that also needs firmware porting and layout review can triple the change cost. It only makes sense at tens of thousands of units a year, or when the original part is also hard to buy.
Consolidate values. Value consolidation is the cheapest engineering lever. A board with 10k pull-ups in both 0402 and 0603, or 100 nF decoupling from three manufacturers, splits its volume across lines that never reach full-reel pricing. Merging them also removes pick-and-place feeder setups at the EMS. A design review is usually enough to approve it, because the electrical value does not change.
Second sources. A pin-compatible second source lowers price and supply risk in the same step. Our cross-reference workflow shows how to check that an alternate is really equivalent before it goes on the approved list.
Worked example: three lines from one industrial controller BOM
The example below applies the levers to three lines of an industrial motor controller built at 8,000 units a year. The part numbers and stock positions are real GlobX catalog entries as of October 2026. The prices in the math are assumptions, labelled as such, because live quotes change weekly.
Line 1: the MCU, lever 4 (timing) instead of lever 8 (downsize). The board uses an ST STM32H743VIT6, a Cortex-M7 in LQFP100 and the most expensive line on the BOM. Downsizing to a smaller H7 would mean porting firmware and redoing validation, which is the wrong first move at 8,000 units. GlobX lists 9,200 units in stock, factory sealed with full labels, date code 21+. That is more than a year of this build, available now rather than at factory lead time. The gotcha: a 21+ date code means the parts may be five years old. Ask for the exact date code, storage conditions and moisture barrier bag status before you price the order.
Line 2: the bus transceiver, lever 6 (grade). The BOM calls for the Nexperia 74HCT245PW-Q100J, the AEC-Q100 automotive variant of an octal bus transceiver in TSSOP-20. This controller goes into factory automation, not vehicles. The standard 74HCT245PW is the same logic function in the same package, without automotive qualification and change control. GlobX lists both with date code 2026. Assume the automotive variant costs EUR 0.06 more. Switching saves EUR 480 a year on this line at almost no qualification cost. Do not switch if the product could be sold into an automotive supply chain later. The qualification is cheap to remove and expensive to add back.
Line 3: passives, lever 5 (consolidate) plus lever 2 (packaging). This line is illustrative rather than a single catalog part. The board has 31 resistor and capacitor lines. Nine of them duplicate a value in a second case size or from a second manufacturer. Merging them leaves 22 lines, and four of the merged lines cross a full reel of 10,000 or 5,000 pieces per year. The saving per part is fractions of a cent. The real gain is at the EMS: nine fewer feeder setups on every build and fewer half-used cut-tape strips in stock.
The result: no engineering change on the most expensive line, one paperwork change on the transceiver, and a design-review change on the passives. That is the usual pattern on a first cost-down pass.
Where BOM cost savings come back as cost
A saving fails when the new part differs from the old one in something the BOM did not record. The usual cases are packaging, moisture sensitivity level and lifecycle status. A cheaper ordering code with a different suffix can be an older die revision, a different MSL rating that the EMS must bake, or a part already marked not recommended for new designs. Each of those costs more later than the saving was worth. Check every changed line against the scrub steps in our BOM scrubbing checklist before it reaches production.
The second failure is treating the prototype BOM as the production BOM. Prototype parts are chosen for availability in tens of pieces: cut tape, catalog stock, sometimes the first part that fitted the footprint. The cost-down pass belongs between pilot and series release, when you know the volume and can get quotes at real quantities.
What a cost-down quote request should contain
A supplier can only quote alternatives and price breaks if the request carries volume and freedom. For each line, send the manufacturer part number, annual volume and release schedule, your current price or target price, and whether alternates are acceptable (none, pin-compatible only, or any equivalent). Add the packaging you need and the oldest date code you will accept. Without annual volume, every quote comes back at a single-lot price.
For a structured first pass, our BOM evaluation service reviews the whole list against stock, lifecycle and price. Or send GlobX the ten most expensive lines of your BOM with annual volumes. We will mark which of the eight levers applies to each line and quote stocked options with date codes within 24 hours.
