Trade-services diligence: cheapest test first

What it is

This is a live deal. The figures below reflect negotiation in progress, not a closed outcome.

HighTide Capital Group buys owner-operated trade-services companies in Arizona and holds them. Each target arrives with a sell-side broker's package: a headline EBITDA figure and an asking price built on it. The engagement covered eleven targets across ten trades.

The job as first scoped was to verify each broker's number. Fowler argued that was the wrong job. A broker's figure is usually arithmetically sound given the seller's books, and of the several figures a set of books can defensibly support, the one presented is usually the largest. Verifying it says nothing about whether the earnings survive a change of owner, which is the only question a buyer is paying for.

The first target made the point: a commercial painting contractor in its 48th year, two brothers retiring, defensible books, a loyal crew and a customer list built on relationships the brothers held personally. Nothing in the package was false. How the owners paid themselves, how many managers the business needed and what the company owed the state were the owners' choices, and none of them transfers to a new owner.

What Common Ground did

• Re-scoped the work before the first target, so every target that followed ran the same three tests in the same order and results compare across trades.
• Ran the liability search first: state and local tax remittance history, pending claims and any owner-carried obligation that conveys with the entity. It is the cheapest test, and it answers what a buyer inherits regardless of price. Arizona's transaction privilege tax is a tax on the business, not a sales tax collected from the customer, and contracting is a taxed classification, so an unremitted balance stays inside a corporation the buyer purchases.
• Ran customer concentration second: revenue by customer, ranked, with the top-two and top-five shares stated, and whose relationships those customers are. A multiple prices earnings, not who the earnings belong to.
• Ran the earnings rebuild last, at two staffing levels: the owners' draw replaced by one market-rate manager, and the fully staffed case with a second manager a new owner needs. Both manager costs were benchmarked to the loaded cost of the seat. Two figures make the choice about how the business will be run visible.
• Took findings to the sellers as findings, in order of size, with two structures to choose from rather than one quiet haircut: a cash path that prices certainty and a partnership path that prices customer retention.
• Ran the full protocol on a second painting contractor the numbers favored. Its ask was set on seven months annualized; rebuilt on filed returns, earnings came in below what the ask needs, the backlog and receivables did not hold up as presented, and a few general contractors held most of the signed work. The recommendation became an offer below the ask, cash priced on the filed years and the backlog paid only as it delivers.
• On an HVAC target the broker cut the ask twice before any test reported. A technology target passed all three tests cleanly and was referred out, because it sat outside the buy box of owner-operated trade-service delivery.
• Graded every exposure finding on a severity scale against the original ask: critical at 35 percent of the ask or more, or a top-two customer share above 50 percent of revenue; high between 15 and 35 percent, or a top-two share of 30 to 50 percent; medium below that. A blank is never a pass.
• Used a standing method for any liability found: buy the assets into a new entity, make a state tax clearance a condition of the close, hold back part of the price with a right of offset against any seller note, and keep the selling corporation funded above a net-worth floor for a period after the close.

Results

• On the first target the liability search found a $472,000 state tax lookback, the single largest number in the deal, before the earnings rebuild began.
• Two customers, both builders and both the owners' own relationships, carried 97 percent of revenue.
• Against the broker's adjusted EBITDA of about $436,000, the rebuild gave $297,000 with one market-rate manager and $181,000 fully staffed, the floor case.
• The ask moved from $1.2 million to $800,000. Under it went two structures: a cash path at $775,000, or a partnership path at $600,000 fixed plus a contingent component tied to customer retention. The counter is still Fowler's to give.
• Across the portfolio, two targets repriced, one passed clean, one had its ask cut by the broker before the tests reported, and seven are still in process. The cheapest test found the biggest number and the most expensive test the smallest; the ordering rule depends only on the cheap test being able to find what the expensive one cannot.
• Installed: the three-test protocol, the severity scale, a findings register per target, a two-level earnings bridge template, and a rule that a finding large enough to move price goes to the seller as a named finding with more than one structure attached.

At a glance