Most portfolio work compresses a catalogue until the buyer can no longer tell which product to ask for. A sequence for finding claims that are narrow, true and worth defending.
The instinct on first contact with a client running several hundred SKUs across four or five families is to compress. Families become platforms, the catalogue is rewritten around a single benefit, and the deck gets shorter. The shortening feels like progress because it is the only visible output.
What tends to be removed in that process is the distinction the buyer was using to choose. A specifier comparing two seal materials is not weighing a brand promise; they are weighing a temperature range and a chemical compatibility list. Flatten both into a line about engineered performance in demanding environments and you have deleted the decision criteria and kept the wrapper.
Consolidation also fails because it is usually run from the inside out. Families exist for reasons of manufacturing history, plant location or acquisition, and grouping by how a company makes things rarely matches how customers shortlist. The grouping that makes sense on a production floor is frequently meaningless to a design engineer choosing a component.
There is one test worth applying before any merge. After the change, can a buyer still work out which product they should be asking for? If the answer is no, the consolidation has cost more than the simplification saved.
These are two claims, aimed at two audiences, resolving two different decisions. Treating them as one is the most common structural fault in an industrial portfolio.
The corporate position answers whether a buyer should deal with this company at all. It is about supply stability, quality systems, engineering support, geographic coverage, lead time behaviour under pressure and whether the firm will still be there in ten years. Procurement and the executive sponsor buy on it.
The product position answers whether a specific part solves a specific problem. It is technical, comparative and narrow, and it is bought by engineers, maintenance leads and specifiers. It has to survive being read next to a competitor's datasheet.
Trouble starts when the corporate claim is pushed down onto product pages, so every technical sheet opens with the same paragraph about innovation and partnership. The reverse fault is just as common: one product family's genuine strength gets elevated into a company-wide claim that the rest of the range cannot carry. Write the corporate position once, keep it to a paragraph, and let it sit as context rather than as the headline on a technical page.
Candidate claims usually fail one of two tests. They are true but unremarkable, which is where quality, reliability and certification claims sit. Or they are remarkable but undefendable, which is where lowest total cost of ownership sits unless someone can produce the arithmetic on request.
A defensible claim almost always has a boundary written into it: a condition, an application, a tolerance, a service commitment. Rebuildable in the field without removing the housing is checkable, and it is either true or false. Superior engineering is neither.
The raw material for the claim is generally sitting in engineering rather than in marketing. Ask why the last dozen competitive jobs were won. Ask what the firm gets asked to quote that competitors decline. Ask the application engineers which problem consumes most of their phone time, because that problem is usually the one the product is genuinely built around.
Then apply the falsification question: what would have to be true for this claim to be wrong. If nobody in the room can answer, the claim is too vague to defend and too vague to repeat. Narrowness feels like surrendering market, but a narrow claim travels accurately through distributors and the client's own sales engineers, and a broad one never does.
A claim engineers approve of can still fail at the point of approval, because the person releasing the money is weighing a different risk. The engineer is assessing whether the part works. The signatory is assessing whether the disruption of changing supplier is justified, and whether they can defend the decision if something goes wrong two years from now.
So the claim needs a two-sided test. The technical audience checks whether it is true and whether it matters. The commercial audience checks whether it is worth the switching cost. A claim that clears one and not the other produces enquiries that stall at the last stage, which is the most expensive place to lose.
The test is not a survey. It is a set of short structured conversations with recent buyers, recent losses and one or two accounts that evaluated the client and stayed where they were. The losses carry more information than the wins, and they are usually easier to get on a call than people expect.
Ask what alternatives were on the list, what nearly stopped the purchase, and what they told their own management in order to get it approved. That last answer is the real positioning, expressed by somebody with nothing to sell. Where distributors or manufacturers' representatives carry competing lines, ask them what they lead with and why; they will normally say plainly.
A position that excludes nothing is not a position. But declining work has to be an operational decision with numbers attached, not a slogan agreed in a workshop and ignored the following Monday.
Sort enquiry types into three groups: work to pursue actively, work to accept but not chase, and work to decline and refer elsewhere. The first two are easy. The third is hard because somebody in the room has a revenue line attached to it.
Make the cost of that third group visible before asking anyone to give it up. The usual argument that lands is engineering load: which enquiry types consume the most application engineering time per unit of margin, and which jobs displace capacity from the work that fits the position. Framed that way, declining stops being a marketing preference and becomes a capacity decision.
Declining in public is also legitimate and frequently underused. Publishing the range boundary, the minimum order, the applications a product is not rated for and the sectors the firm does not serve filters enquiries before they consume a person's afternoon.
Rebranding an acquisition quickly is the default and it is often wrong. The question is not how much equity the old name is said to hold, but where that name appears in the customer's buying process.
If the name is written on drawings, sits in approved-vendor lists, appears in maintenance manuals or has been used in a standards or specification submission, removing it creates re-specification work for the customer. Re-specification is precisely where an incumbent gets displaced, because it reopens a decision that had been closed. That risk is usually larger than the cost of running two names.
So the working rule for is to keep the name where it carries specification weight and retire it where it only carries sales history. An endorsed structure, with the acquired name followed by the parent, is the common middle path. Its value is practical rather than architectural: it gives procurement a way to update their records without re-approving the part.
Set a review date rather than a permanent answer. Specification weight decays as the installed base turns over, and a name that must be kept today may be retirable in three years.
The work above can be compressed into a single working session, provided the preparation is done first. Preparation means three artefacts: a map of the catalogue against revenue and margin, a win and loss list covering the last eighteen months, and a written list of the questions the application engineers are actually asked. Without those, the day becomes an opinion exchange.
Attendance matters more than agenda. You need at least one application or design engineer, one person who carries a number, and one person senior enough to authorise a decline. Marketing alone cannot produce a defensible claim, because marketing does not hold the evidence.
Run it in this order, and keep each step timed.
The output is not finished messaging. It is a set of claims with named evidence gaps, and the following weeks are spent closing those gaps: pulling the test data, writing the application note, getting the comparison approved by someone who can be held to it.
Resist the pressure to rewrite the website before that work is done. A claim published without its supporting evidence gets challenged by the first competent engineer who reads it, and the client then retreats to generic language, which is how portfolios end up flat in the first place.
Judge the result on a narrow criterion. Six months later, can a distributor, a field sales engineer and the marketing manager each state the claim for a given family in roughly the same words, and point to the evidence behind it. If they can, the position is holding. If they cannot, it was too broad to carry.
Key takeaways
The count is the wrong unit. Position at the level customers shortlist at, which is usually the family or the application, not the part number. Individual SKUs need accurate technical data and clear selection logic rather than positioning of their own.
Decide by where each name appears in the customer's buying process. A name written into drawings, approved-vendor lists or standards submissions carries specification weight and is expensive to remove. A name that only carries sales history can usually be retired into an endorsed structure.
A boundary and a source of evidence. A claim with a stated condition, application or tolerance can be checked and either confirmed or refuted. A claim with no boundary cannot be defended when a competitor's engineer reads it next to their own datasheet.
Convert it from a marketing preference into a capacity argument. Show which enquiry types consume the most application engineering time per unit of margin, and which displace work that fits the position. Then agree the decline list with someone senior enough to hold it.
