Version 2 Is a Business Decision Wearing an Engineering Costume

Once a product is shipping, pressure to build the next one arrives from every direction. Customers want the feature you cut. Support wants the failure mode gone. Sales wants a reason to call old accounts. Engineering wants to fix the compromises they still remember. All legitimate, and none of them alone justifies six months and a tooling budget.

The question that decides it is narrow: does the change move a number that matters — return rate, unit cost, attach rate, win rate, reorder rate — enough to pay back the investment inside a year or two of shipping? A version 2 adding features nobody was blocked on just spends the profit from version 1.

The Field Signals That Genuinely Justify a Second Generation

After twelve to eighteen months of real shipments you have data no amount of pre-launch research could produce. The signals worth acting on:

  • A concentrated warranty failure. When one part accounts for most of your RMAs, the fix is worth doing regardless of anything else. A 4 percent return rate on a $150 product with a $40 handling and replacement cost is $240 of loss per hundred units sold, and it compounds through reviews.
  • A repeated lost-deal reason. If sales calls die on the same missing capability three times a quarter, that capability has a measurable dollar value. If the reasons are scattered, you have a positioning problem instead.
  • Support load per unit. Track tickets per hundred units in the field. A setup step that generates calls costs real money and is usually cheap to design out.
  • Component obsolescence. A microcontroller or sensor going end-of-life forces a board revision anyway, and that is the moment to bundle other changes. The decision tree is in a component went end-of-life.
  • Cost structure that blocks a channel. If a retail account requires a lower unit cost than the current design allows, redesigning for cost has a known payoff, and the method is value engineering rather than new features.
  • Observed misuse. Users consistently doing something you did not design for is a product-definition finding, and it often points at a variant rather than a replacement.

Signals that do not justify version 2: a single loud customer, a competitor's spec sheet, and internal boredom.

Not Every Change Needs a New Generation

There are four escalating levels of change, and choosing the smallest one that solves the problem is most of the skill.

  1. Firmware update. Behavior, logic, and often performance, at effectively zero unit cost if you already have OTA update capability. If you do not, adding it should be near the top of your version 2 list.
  2. Running change. A supplier swap, a tolerance adjustment, a fastener change. Handled through the engineering change order process with no announcement and no new SKU.
  3. Minor revision. Steel-safe tooling changes, a new label, a firmware plus hardware pairing. Sold as the same product.
  4. New generation. New tooling, new part numbers, new packaging, a marketing moment, and a plan for the old inventory.

A full teardown-and-restart belongs in a different conversation entirely, covered in redesigning an existing product.

What Tooling Changes Actually Cost

The single number that most often decides scope is how much of the existing mold survives. Steel can be removed cheaply and added expensively.

  • Steel-safe change — adding material to the part, which means removing steel from the cavity. Typically $500 to $3,000 and one to two weeks, plus a sample run.
  • Insert replacement — swapping a modular insert to change a feature. $2,000 to $12,000 depending on size and complexity, three to five weeks.
  • Welding and remachining — adding steel back into a cavity. Possible, sometimes invisible, sometimes leaves a witness line on cosmetic surfaces. $3,000 to $15,000, and it consumes some of the tool's remaining life.
  • New tool — a fresh multi-cavity production mold for a mid-size housing commonly runs $18,000 to $60,000 with a twelve-to-sixteen-week lead time.

Design version 2 backward from that list. If a feature can be achieved by adding plastic rather than removing it, the tooling cost drops by an order of magnitude. Check the remaining shot count on the existing tools before you plan anything, using the criteria in how many shots a mold lasts, because a tool near end of life changes the math entirely: if it needs replacing anyway, the incremental cost of redesigning the part is small.

Add the invisible costs. New photography, updated manuals and packaging artwork, certification retesting if the enclosure or radio environment changed, retraining support, and updating the bill of materials and every downstream document. These routinely add $15,000 to $40,000 on top of engineering and tooling.

Backward Compatibility Is a Product Decision

Two compatibility questions decide how much goodwill version 2 costs you.

Accessories and consumables

If customers own mounts, batteries, blades, filters, or cartridges, breaking the interface converts loyal owners into people who feel punished. In a razor-and-blade model it can be fatal. Freeze the mating interface — bolt pattern, port geometry, cartridge envelope — and treat it as a published specification. Everything behind it can change freely.

Firmware and app

Both generations will be in the field for years, so plan explicitly: a hardware revision byte the firmware can read, one app that supports both, a protocol that negotiates capability instead of assuming it, and an OTA path so a v1 unit pulled from a warehouse years later updates itself on first connection. What a v1 device does when it meets a v2 app is a five-minute design conversation and a three-week post-launch emergency.

Sequencing the Transition

Run down v1 inventory before announcing v2 unless you plan to sell both, because announcing early stops sales of what is still on the shelf. Decide whether v2 replaces v1 or sits above it as a premium tier — keeping both is legitimate when the tooling is paid for and the SKUs serve different buyers. Give the channel a defined transition window, honor v1 warranties with parts you have actually stocked, and validate the new build with a proper pilot production run. Second generations fail in production for the same reasons first ones do.

Deciding What Version 2 Should Be

Projects House works with companies that already have a product in the field: reviewing field data, scoping what the existing tooling can absorb, and engineering a second generation that preserves the installed base. Send your current product, your warranty and support data, and what you want to change through our contact form.