Run a sourcing process in any reasonably defined UK private sector and ask three independent deal teams to produce a long-list of targets. The overlap between the three lists will be significant: perhaps ten to fifteen companies appear on all three, another twenty appear on at least two. The teams did not copy each other. They used different databases, different networks, different approaches. But they converged on substantially the same set of names.
This convergence is not coincidence and it is not a sign that the sourcing was done well. It is a symptom of a structural problem in how deal teams build target universes: they draw from the same visible layer of the market, and below that layer there is a second, larger population that none of the standard methods systematically surfaces.
Understanding why the convergence happens, and what the less visible population looks like, is the starting point for thinking about how to source differently.
Why the Same Names Appear
The standard methods for building a target universe share a common selection mechanism. They start from data sources that index companies by their public visibility, their size, or their classification, and they apply filters to reduce that universe to a manageable list.
Database products like Fame and Orbis index companies by Companies House filings and apply standardised size estimates. They are well-suited to finding companies that file full accounts (because those contain the revenue and profit data that databases use for filtering) and companies that self-classify accurately under recognisable SIC codes. These companies tend to be larger, professionally managed, and often already known to advisers.
Trade press and industry association membership lists surface the companies that are most active in their industry association, most likely to have been covered in trade publications, and most likely to sponsor or speak at industry events. These are typically the sector's visible participants: not necessarily the best businesses, but the ones most invested in sector visibility.
Adviser referrals surface companies that have already been in or adjacent to formal processes, or whose owners have explicitly started having conversations about succession or exit. These companies are, by definition, closer to the visible end of the market. A company whose owner has spoken to three advisers in the past year will appear in referral networks. A company whose owner has not yet had those conversations will not.
Each of these methods, applied diligently, produces a list of thirty to fifty companies in a typical UK mid-market sector. The problem is that these lists overlap substantially across different teams using the same methods, and they collectively describe the same visible portion of the universe.
What the Visible Layer Looks Like
The companies that appear consistently on independently-produced long-lists share several observable characteristics. They are typically larger within their sub-segment: because size correlates with database indexing quality, professional management (which correlates with public visibility), and adviser coverage. They are often PE-backed or have previously had PE ownership: because institutional ownership creates relationships with the intermediary community that generates referral flow. They have typically been involved in at least one prior transaction, even if as a seller in a partial deal: because transaction history creates adviser relationships and database coverage.
None of this makes these companies bad targets. The visible names are often genuinely good businesses. The problem is that selecting only from this population means competing for the same assets as every other buyer who used the same methods. Competitive tension in a process concentrates on a small set of names. Deal teams that want to develop a proprietary position in a sector need to find companies that are genuinely less competed for, which means finding them before they become visible through the standard channels.
The Less Visible Population
Below the visible layer there is a larger population of companies that meet reasonable size and quality criteria but are structurally absent from the standard lists. These companies share characteristics that are the inverse of the visible layer.
They are typically founder-owned, with no institutional ownership and no prior transaction history. Their owners have not spoken to advisers because they have not yet decided to sell, or have decided not to engage a process. They may have modest online presence and limited trade press coverage. They may self-classify under SIC codes that are imprecise or outdated. Their accounts are abbreviated, so no revenue figure appears in the database.
These companies are not invisible in an absolute sense. They all file at Companies House. Many of them advertise externally. Their directors appear in the public record. Their balance sheets tell a story about their size and financial health. But extracting them from the public record requires a different method than filtering a commercial database: it requires structured extraction from primary sources with a methodology that does not select for visibility.
This is the population that tends to produce better deal outcomes for buyers willing to approach before a formal process, for a straightforward reason: there is less competition for them. A company that no other buyer has identified and approached is a company where the first serious, credible approach has pricing power that the same company in a competitive process would not have.
The Structural Reason This Is Hard to Fix
The convergence problem persists because the incentives that produce it are rational at the level of individual deal teams. Database products and adviser relationships are genuinely faster ways to build an initial universe than structured public record extraction. The bias toward visible names is a rational response to resource constraints: it is cheaper and faster to identify visible companies than invisible ones, so under time pressure, deal teams rationally concentrate effort on the visible layer.
This rational individual decision produces a collective outcome that is suboptimal for the market. Every buyer concentrating on the same visible layer creates competition for those assets, which raises prices. The less visible layer remains undisourced. The gap between visible and less visible target quality (which is not correlated with visibility) persists.
Closing that gap requires systematically investing in the structured extraction work that surfaces the less visible population. This is not something that pays off immediately. The companies in the less visible population are not ready to transact today. What the work produces is a pipeline: companies identified eighteen months before they become visible through standard channels, with a proprietary contact established before competitive processes begin.
What Differentiated Sourcing Looks Like in Practice
The practical difference between standard sourcing and differentiated sourcing is most visible in the composition of the target list, not in the methodology itself. A list generated from database screening and adviser referrals will look like a list of the sector's known participants. A list generated from structured public record extraction will include a substantial proportion of companies that do not appear in any database product and have never been to market.
For the less visible companies, the evidence base for each is different from what a database screen produces. There is no revenue figure from a disclosed account. Instead, there is a triangulated size estimate from balance sheet proxies and employment data, a hiring signal profile that suggests current trajectory, and an ownership structure read from confirmation statements that characterises the likely decision-making dynamic around any future transaction.
This is what Thema produces when we map a sector. The output is not a list of the companies that every other buyer working the sector already knows. It is a comprehensive view of the sector's structure that gives equal analytical weight to the visible and less visible populations, ranked by the signals that predict transaction readiness, not by size or adviser coverage. The twelve companies on everyone's list will appear in the map. The forty companies that are not on anyone's list will appear there too. Which set is more interesting depends on what you are trying to do.