Why industry codes miss your best acquisition targets
By
Lisa
Most longlists still start the same way: pick a NAICS, SIC, or NACE code, pull every company registered under it, and start filtering. It feels systematic. It is also how a lot of good targets never make it onto the list.
Industry codes were built for statistics, not deal sourcing. They group companies by what a government classification scheme says a business does, not by what actually makes it a fit as an acquisition target. For corporate development, M&A advisory, and private equity teams, that gap shows up as missed targets, bloated longlists, and hours spent manually filtering out companies that never belonged on the list in the first place.
In this article you'll learn:
- Why NAICS, SIC, and NACE codes systematically miss acquisition targets, especially in niche and emerging sectors
- What the data shows about how often industry codes misclassify companies
- Which dimensions to search by instead of, or alongside, industry codes
- How a code-only longlist compares to a dimension-based one in practice
- How to combine industry codes with other search criteria without losing speed
Why industry codes miss your best acquisition targets
The problem: one code, many different businesses
Industry classification systems are hierarchical by design. NACE Rev. 2, the EU system, has 21 sections, 88 divisions, 272 groups, and 615 classes. That sounds granular until you look at what ends up inside a single class.
In the UK, SIC code 82990, “other business support services not elsewhere classified,” covers Amazon UK Services, Google UK, Santander UK, and Facebook UK, four companies with almost nothing in common beyond having been assigned the same catch-all code. A search built around that code returns all of them, plus every other company that happened to land in the same bucket.
The codes are also slow to change. The UK's SIC system was last meaningfully revised in 2007, and the gap before the next revision, SIC 2026, is the longest in its history at 19 years. New industries can generate revenue for a decade before a dedicated code exists for them. Until then, they sit in generic or “not elsewhere classified” categories, the same categories that absorb companies a classification system simply doesn't know how to describe.
Codes also drift from reality over time. Companies choose their own code at registration, and in most jurisdictions nothing forces them to update it as the business evolves. The UK's Office for National Statistics found that SIC codes mismatched a company's actual reported activity for 11.5% of companies at the most detailed classification level, and 8.7% at the broader division level.
Why this matters more for niche and cross-border searches
For a generalist search, a bloated or slightly outdated code might just mean extra filtering. For deal teams working in a specific niche, cross-border, or in an emerging category, it means something more costly: targets that never show up at all.
A company operating in a niche segment with no dedicated code gets filed under “other,” alongside dozens of unrelated businesses. A company that updated its product or business model since incorporation is still sitting under its original code. A target in a different country will be classified under a different system entirely, since NAICS (US), SIC (UK), NACE (EU), and national systems such as Germany's WZ do not map onto each other one to one. None of this means the target doesn't exist. It means a code-based search cannot see it.
What a code-only longlist misses in practice
This is a familiar pattern for teams that query databases one industry code at a time. As Harald Friedrich, Vice President at VINCI Energies, put it: “Instead of querying databases individually by NACE code, StrategyBridgeAI delivers structured results in a short time, including matches the team hadn't had on its radar.”
The phrase worth noting is “hadn't had on its radar.” Those aren't targets the team evaluated and rejected. They're targets a code-based search structurally could not surface in the first place.
What to search by instead: dimensions that define a target, not a code
A more reliable starting point is to define the target by the dimensions that actually matter for the deal: geography, product focus, market segment, business model, customer base, size, and growth profile. None of these require a classification code to express, and all of them can be combined and adjusted as the search evolves.
Industry codes vs. dimension-based search
| Industry code search (NAICS / SIC / NACE) | Dimension-based search | |
|---|---|---|
| What groups companies | A fixed numeric category assigned at registration | Any combination of criteria: product, geography, business model, market segment |
| How current it is | Revised roughly once a decade; some codes are 15+ years old | Updated continuously as new company data becomes available |
| New or niche industries | Often absorbed into generic or “not elsewhere classified” categories | Searchable from day one by what the company actually does |
| Cross-border consistency | NAICS, SIC, NACE, and national systems such as Germany's WZ don't map 1:1 | The same search dimensions apply across markets |
| Typical output | A list of every company sharing a code, regardless of fit | A list of companies matching the actual target profile |
This doesn't make industry codes useless. They're still a fast way to build a standard peer group or satisfy a reporting requirement. The point is that they shouldn't be the only filter standing between a deal team and a target that fits.
How StrategyBridgeAI finds targets beyond industry codes
StrategyBridgeAI's Longlist is built around this gap. You simply describe the target you're looking for in a chat interface, in plain language, geography, product focus, market segment, business model, whatever defines it, and the platform searches on that. There's no code to look up first, and no risk of forgetting a filter, since you're not assembling one from a checklist. Structured filters are still available alongside the chat, organized into four groups: company, location, financials, and ownership, with example criteria such as sector or listing status, region or country, revenue or EBITDA, and private-equity backed or state-owned, for whenever you want to narrow results further. But the starting point is always a plain-language search, not a classification code, which is what surfaces small and mid-sized companies and niche providers that a purely code-bound database would not return.
Standard codes aren't excluded. NACE, SIC, GICS, and country-specific systems such as Germany's WZ are all supported where they're useful, for example to anchor a peer group or meet a reporting standard. The difference is that they're one input among several, not the single gate a target has to pass through to be found.
A longlist doesn't have to stop at the first result either. You can extend it with your own custom columns and enrich it with additional information pulled directly from the web, right inside the interface, instead of exporting it to look up context elsewhere. From there, one click on a company takes you straight into its full profile inside the platform, giving you what you need for a first screening without leaving the workflow.
Longlist draws on StrategyBridgeAI's database of roughly 50 million public and private companies across more than 100 countries, so a dimension-based search isn't limited to a single market's classification system.
Teams report cutting the overall research and screening effort behind a longlist from a historical range of 20 to 120 hours down to under five, a difference that mostly comes from not having to manually chase down the targets a code-based search missed in the first place.
Longlist isn't limited to building a list from scratch, either. Upload a longlist you've already built, in whatever format you have it, and the platform matches it against verified company records and fills in the gaps. The same search logic applies beyond acquisition targets too, to finding strategic buyers and financial investors, contact details included. And whichever workflow you're in, the result exports directly to Excel.
A practical approach: combine codes and dimensions
In practice, start with the dimension-based, semantic search: describe the target directly, and the first pass already returns companies that are genuinely relevant, not a bucket of companies that happen to share a code. That's usually where the most interesting targets surface, the niche player with no dedicated code, the company that outgrew its original classification, the cross-border target filed under a different system entirely.
Industry codes still have a role after that: anchoring a standard peer group, satisfying a reporting requirement, or mapping results onto a classification stakeholders already recognize. For a longlist that needs to hold up to scrutiny in an investment committee or sell-side process, adding that layer isn't optional, even once the dimension-based search has already done the work of finding the right companies.
Frequently asked questions
What's wrong with using NAICS or SIC codes to build an M&A longlist?+
Industry codes group companies by a classification assigned at registration, not by what actually makes a company a fit as an acquisition target. Codes can be broad (bundling unrelated businesses together), outdated (some systems go over a decade between revisions), or simply wrong, since companies rarely update their code as the business changes. Used alone, they let through companies that don't fit and miss others that do.
How do I find acquisition targets that don't have the right industry code?+
Search by the dimensions that define the target instead of, or in addition to, a classification code: geography, product focus, market segment, business model, customer base, and size. These criteria can be combined and refined as the search evolves, and they surface niche or emerging businesses that haven't been assigned, or have outgrown, a specific code.
Can I combine industry codes with other search criteria in a longlist tool?+
Yes. Industry codes such as NACE, SIC, GICS, and country-specific systems like Germany's WZ remain useful for anchoring a standard peer group or meeting a reporting requirement. A tool like StrategyBridgeAI's Longlist supports standard codes alongside dimension-based filters, so a search isn't limited to one or the other.
Why do different countries use different industry classification systems?+
Each system, NAICS in the US, SIC in the UK, NACE in the EU, Germany's WZ, and others, was developed independently for national statistical purposes. They don't map onto each other one to one, which is why a cross-border target search based purely on codes often misses targets that a different country's system would classify differently, or not at all.
How often are industry classification codes updated?+
Revisions typically happen roughly once a decade, and some gaps have been longer. The UK's SIC system, for example, went 19 years between its two most recent major revisions, the longest gap in its history. New or fast-moving industries can exist for years before a dedicated code is introduced for them.
More from this category
Explore related insights from StrategyBridgeAI.
Who will buy your client's company? What 1.6 million acquisitions tell sell-side advisors
10% of acquirers make 56% of all deals, financial investors 36%. What 1.6 million acquisitions tell sell-side advisors about building a stronger buyer universe.
M&A outlook 2027: why next year's buyers are already active today
85% of companies with more than 100 acquisitions were active in the past two years. Why next year's buyer landscape is visible today, and where AI still leaves a gap in M&A.
How to score and prioritize acquisition targets on your M&A longlist
A practical scoring framework for ranking acquisition targets on an M&A longlist, so deal teams spend outreach time on the targets most likely to close.