$37.02
$88.01
bear $37.02
base $63.45
bull $88.01
margin of safety $32.00
market price* $65.35
What the business is. Tennant, founded in Minneapolis in 1870, designs, builds and services commercial and industrial floor-cleaning machines: ride-on and walk-behind scrubbers and sweepers, autonomous robotic scrubbers, and the parts, consumables and service that keep its installed base running. Brands include Tennant, Nobles, IPC (Italy, bought in April 2017 for about $354 million), Gaomei (China, 2019) and TCS (Europe, 2024). Sales in 2025 were $1.20 billion: Americas $792 million, Europe/Middle East/Africa $335 million, Asia-Pacific $77 million. Customers are contract cleaners, retailers, warehouses, factories and facility managers; demand follows business investment and the commercial-construction and industrial cycle. How we measured earnings. We start from operating cash flow and strip out working-capital swings, which are very large here (inventory alone moved +$56 million in 2021 and +$68 million in 2022, then partly reversed). Share-based pay and bad-debt and returns provisions are charged as real costs. One judgement matters: Tennant is replacing its ERP (business software) system and capitalizes the implementation cost as an "other asset", so the cash ($23.3 million in 2024, $30.6 million in 2025) runs through operating cash flow and would vanish when working capital is stripped out. We treat it as capital spending. Amortization of acquired intangibles ($13.7 million in 2025, mostly IPC) is added back. IPC's brands and dealer relationships are still selling, so the base case models no decay; the bear case carries erosion through a permanently lower earnings level. Owner earnings per share: 2016 $2.06, 2017 $1.46 (IPC acquisition year), 2018 $3.46, 2019 $3.06, 2020 $2.97, 2021 $4.77, 2022 $3.84, 2023 $6.19, 2024 $4.53, 2025 $3.00. We anchor on 2018-2025, the business as it has existed since IPC. That window includes the 2020 shutdown, the 2022 supply-chain squeeze and the 2025 ERP disruption, as well as the 2023 peak, when a backlog was being worked down. The median is $3.65 per share. Adding back the cash interest paid gives the cash the business makes before its lenders: a median of about $83 million a year. What is happening now. In the first week of November 2025 the new ERP system went live in North America. The company lost three weeks of machine order entry and parts shipping. In January 2026 it shut its factories for two weeks to take a full physical inventory. Tariffs added since mid-2025 are squeezing margins. Sales in the first half of 2026 rose 2% to $622 million, but operating income fell to $20.8 million from $50.2 million, and the second quarter was still down by half year on year. Business cash flow fell about 32% in 2025, from $100 million to $68 million. For 2026 we assume about $62 million, roughly 9% lower. We assume a smaller decline than 2025's because about $35 million had already been earned in the first half of 2026, against roughly $42 million a year earlier, and because 2025 carried both the largest year of ERP spending and the go-live disruption. Base case. About $62 million in 2026 and $74 million in 2027, back to the $83 million median by 2028 as the ERP settles and price increases catch up with tariffs, then flat. We credit no growth: revenue grew only about 1% a year from 2018 to 2025 even with acquisitions, and owner earnings show no trend. Earnings continue past ten years because the installed base keeps buying parts, consumables and service through a direct service network, as it has for more than a century. The business is worth about $1.40 billion. After net debt of $312 million, the equity is about $1.09 billion, or $63 per share. Bear case. ERP problems recur as the system is rolled out to Europe and Asia, and tariffs stick, holding 2026-2027 near $55 million. Recovery to $68 million in 2028 is then cut short by a recession of the kind this industry went through in 2008-2009 (Tennant's sales fell about 12% in 2020 alone). Cash flow drops to about $20 million in 2029, recovers to $40 million in 2030, and settles permanently at $60 million as robotic and lower-cost competitors and higher reinvestment hold margins down. At the trough, debt would exceed the lenders' 3.75x debt-to-EBITDA limit, forcing an amendment at a higher spread (we charge about $25 million), a halt to buybacks and pressure on the dividend. The business is worth about $973 million and owes about $337 million including that cost, leaving about $636 million, or $37 per share. Bull case. Margins return to the 2021-2024 level, about $100 million of cash a year by 2028. After that, earnings grow 1.5% a year through 2035, based on the company's record of annual price increases and small bolt-on acquisitions (Gaomei, TCS), then hold flat. The business is worth about $1.82 billion; less $312 million of net debt, that leaves about $1.51 billion, or $88 per share. Debt. On June 30, 2026, borrowings were $358.9 million, up from $273.6 million at year-end, mainly because the company repurchased $60.5 million of stock in the first half while earnings were falling. Cash was $76.9 million, of which we treat about $47 million as surplus. The debt is drawn mainly under a $650 million secured revolving credit facility, with JPMorgan as agent. It runs to August 7, 2029 at floating rates (SOFR plus 1.25-2.0%). Its covenants cap debt at 3.75x EBITDA, require interest coverage of at least 3.0x, and limit buybacks and dividends once leverage passes 2.5x. With trailing operating income depressed, leverage is much closer to those limits than it was a year ago. The 2029 maturity and the covenants are what matter most for the bear case. What we could not reconcile. Depreciation ($45 million in 2025) runs about twice recorded capital spending ($21.7 million), and net property and equipment has not shrunk the way that gap implies. The filings do not show where the extra additions come from (currency, acquisitions or equipment moved from inventory into service). If part of it is real reinvestment that runs through inventory, owner earnings are overstated by perhaps $10-15 million a year. The bear case's lower permanent level is meant to absorb that. Margin-of-safety price. We are reasonably, not highly, confident in the earnings level. Eight years of results since IPC range from $68 million to $133 million of business cash flow. The last three quarters have been weak, and the causes (a botched software cutover and tariffs) are identifiable but not yet shown to be over. The worst plausible outcome is a recession arriving before the ERP is stable, with leverage already near covenant limits. The structure would survive it: no customer concentration, plants in the US, the Netherlands, Italy, Brazil and China, a large recurring parts and service stream, and positive cash flow even at the trough. Owners would lose buybacks and possibly part of the dividend; dilution is unlikely. The things we are most likely to be wrong about: whether margins recover to the post-2018 median (if they settle near 2025 levels, the shares are worth roughly $49); whether the European and Asian ERP rollouts repeat the North American disruption; and the unreconciled depreciation gap (worth about $10-13 per share). The price assumes a three-to-four-year holding period for the recovery to show up, with a dividend of about $1.20 a share a year. From $32, a recovery to the median earning power over that period implies close to 20% a year including dividends, and the bear case value is above that price. The margin-of-safety price is $32. Earlier published figures for this company were prepared on a different basis and are not comparable.