A KeePer coating is a thin glass film that goes onto car paint. The cheapest one, Crystal KeePer, takes a couple of hours and costs ¥18,200, about $117, on a small car. The most expensive, W Diamond KeePer, takes most of a day and costs ¥108,000, about $690, on a large one. What the customer is buying is a brand they recognise and trust, and a system that produces the same result in every shop allowed to use the name.
That is the idea behind this company. It built the brand and the method, then let other people’s shops do the work.
The shares trade at ¥3,650, which values KeePer Technical Laboratory (TSE: 6036, KeePer技研) at ¥99.6bn, or about $638m at ¥156 to the dollar. Revenue in the twelve months to June 2026 was ¥25.96bn, up 12.4%.
[Video: KeePer’s own commercial for a KeePer LABO. ]
What KeePer sells
KeePer designs the coatings and has them made under contract by SONAX, a German manufacturer. In May 2024 the two replaced an open-ended supply deal with an exclusive worldwide agreement running ten years.
That chemistry goes mostly into two markets. The aftermarket, meaning shops that work on cars people already own, is 60% of product revenue and runs mostly through gas stations. New car dealers are most of the rest and growing quickly, because KeePer is an approved factory option at Toyota, Honda, Subaru, Volvo and Mercedes-Benz, and more than nine in ten new cars in Japan still leave the dealer without any coating at all.
KeePer is in that aftermarket itself as well. KeePer LABO is its own format, a specialist store for car coating and washing, with several bays under one roof and staff who do nothing else all day. There were 187 of them at the end of August. Selling the chemistry and running your own LABOs are completely different businesses, and the last twelve months show how different.
The LABO chain sells 23% more than the product side and earns 46% less from it. It also eats the capital. In the year to June 2025, the only period split by segment, the product business used ¥30m and the chain ¥1.97bn. On the assets each side employs, the product business earned a 112% return and the chain 25%.
So the better half of this company is the half that sells to shops it does not own. The problem is that most of the shops are gas stations, and gas stations are closing. Which is why the interesting part of this company is what it started two years ago. It stopped building LABOs itself and began franchising them, turning the expensive half into a licence business too.
The 6,661 shops it does not own
A shop that wants to sell KeePer coatings sends staff to one of 22 training centres to sit an exam. If they pass and the site is good enough, it becomes a certified KeePer Pro Shop, with no fee to join and none to stay. By June 2025 there were 6,661 of them, about 23% of every gas station in Japan. KeePer pays none of their rent, staff or electricity. All it controls is who may use the name. KeePer says in its own risk section that no competitor puts product development, wholesale, a retail chain and a training school in one company, and I could not find one that does.
That number has stopped growing, and the company expects it to fall. Japan has more than 29,000 gas stations and their main product is going away, because cars use less petrol, people drive shorter distances, and electric cars will speed both up. The annual report says it plainly: a decline in the number of Pro Shops is unavoidable. But it also says why it is not worried. Making cars beautiful will exist as long as cars exist, it argues, even once the age of the electric car arrives, and the gas station is still the place drivers visit most often.
So picture the station owner. A building on a main road, trained staff, regular customers, and a product that is disappearing. What he needs is a second business for the same building.
Where the growth comes from
Until two years ago KeePer built every LABO itself. That changed in August 2024, when it signed a partnership with Idemitsu, one of Japan’s big oil companies, and opened franchise recruitment three months later. About fifty companies with more than a hundred sites between them have applied since, mostly station operators who already employ technicians.
A LABO is the tier above a Pro Shop. At a Pro Shop the station’s staff do the coating in between serving fuel, so the finish and the waiting time vary from site to site. About half of first-time LABO customers go to a nearby Pro Shop for the second coating.
An owned LABO takes in about ¥8.1m a month and every yen of it is KeePer’s revenue. It costs about ¥70m to build, and KeePer pays every wage inside it. There are 150 of these.
A franchised LABO takes in about ¥5.6m a month and KeePer receives ¥1.2m of it, a take rate of 21.6%. Part of that is chemicals rather than a pure fee, so it does not all drop to profit, but the capital really is nothing. It costs KeePer nothing to build, and the partner pays every wage. There are 37 of these, and the revenue they send KeePer grew 91% in August.
For now the two run in parallel, and last year KeePer opened thirteen of its own and fifteen franchises. But the direction is set. Pro Shop numbers are going to fall, an owned LABO costs ¥70m of KeePer’s own money, and the franchise is the only one of the three that grows without capital.
Why profit fell last year
Operating profit for the twelve months to June 2026 fell 2.6% while revenue grew 12.4%, and the margin dropped from 30.7% to 26.6%. For a company most investors treat as a compounder, that looks like something breaking.
It isn’t. In November 2025 KeePer sold a stake in another listed Japanese car care company and made ¥6.8bn on it. In February it said it would spend ¥1,031m of that on a television campaign and a bonus for every employee.
The ¥6.8bn does not count as operating profit, because selling shares is not KeePer’s business. The ¥1,031m does count as an operating cost. So the gain never showed up in operating profit, but the spending did.
Add the ¥1,031m back and profit rose 11.9%, with the margin near 30.6%. The company runs the same calculation in its half-year report, but only for the first half, so the full year is my own work.
Guidance for June 2027 is ¥9,012m. Against the reported number that is a 30% jump. Against the normalised number it is 13.4%, on revenue growth of 13.2%, with the margin unchanged. KeePer is not promising a recovery, it is promising not to spend a windfall twice.
One year is the wrong length to judge this company by, so here is the whole record.
Revenue has compounded at about 15% a year for a decade and profit faster, because the operating margin went from 12% to 30% along the way.
What worries me
The LABOs are getting more expensive to run. Their revenue grew 10.5% last year while the cost of running them grew 23.6%, and the margin inside the chain fell from 46.8% to 40.5%. Most of that is wages, and in a country running out of workers that pressure will not go away. KeePer also buys its chemicals from Germany, so a weak yen does not help, but wages are far the bigger number. Cash has been tighter too: operating cash flow covered 87% of operating profit two years ago, 82% the year after and 64% over the last twelve months.
There is a counterweight. KeePer raised its prices in June, and by July almost every job was being done at the new price. At stores open more than a year the average spend per car rose 7.6% in July and 10.4% in August. Visits were flat in July and down 4% in August, when heavy rain kept the cheap car wash customers away. Two months is not proof, but the increase stuck.
The LABO side is also the weaker moat. Sales at LABOs open more than a year grew 2%, margins fall in hot summers, and the only real barrier is a trained technician. Also one distributor, ENEOS Trading, is 11% of revenue and 30% of what KeePer is owed.
And the founder is 74. His holding company owns 21% and a car dealer group another 17%, so the float is thin. Meanwhile ¥13.3bn of net cash sits idle, 13% of the market value, and the last buyback was in 2021. Some of the discount in the share price is sitting right here.
The annual report lands on 25 September with the new segment split, the Pro Shop count and this year’s store plan. That is when I will know more.
What the stock costs
At ¥3,650 the enterprise value, meaning the shares plus debt minus cash, is about ¥86bn. That is 10.9 times the last twelve months of normalised operating profit and 9.6 times what the company expects this year. The dividend yields 2.7%.
Work backwards from that price. Take this year's guided operating profit, tax it at the 31% KeePer actually pays, and the market is paying 14 times what is left. If you want 10% a year on your own money, that is a company the market expects to grow at about 3% a year forever. It grew 12.4% last year and is guiding for 13% this year. The shares have also de-rated on every measure: the market paid between 25 and 34 times earnings every year from 2021 to 2024 and pays 16 times now. One broker follows the company, Tokai Tokyo Securities, and it rates the shares neutral with a target of ¥3,070, about 16% below the price, cut six times since August 2023 when it stood at ¥7,480.
My own five years look like this. The margin starts at the 30.6% the business made last year once the one-off spending is stripped out, and drifts up to about 32%, because the franchise revenue inside the group grows quicker than everything else and carries far less cost. Revenue compounds at about 10% a year and profit at about 11.5%.
The bear case is simple. Gas stations close faster than KeePer converts them, wages keep eating the LABO margin, and you own an ordinary retailer at a fair price.
One last thing. Everything above comes out of the filings and the public sources listed below. So if you live in Japan, or you know this industry better than I do, I would really like to hear from you, in the comments or by private message. Corrections and criticism are very welcome.
Sources. Everything here comes from KeePer Technical Laboratory's own filings on EDINET: the annual reports for FY6/2016 to FY6/2025, the half-year reports to December 2024 and December 2025, the full-year results release of 14 August 2026, the forecast revision of 16 February 2026, the filings on the Soft99 disposal, and the monthly sales reports through August 2026. Twelve-month segment figures are FY6/2025 plus the December 2025 half minus the December 2024 half, which is my own construction, as is the full-year normalisation. Coating prices are KeePer's own published price list and are the only figures here that are not from a filing. Analyst coverage is from IFIS and Nikkei Yosoku. Share price ¥3,650 as of 10 September 2026, ¥156.25 to the dollar.
How I work. I read the Japanese filings myself. I have never visited a KeePer store, so everything here comes out of documents. AI helps me with translation and with checking the arithmetic. Every judgement, model and conclusion here is mine.
Disclosure. Assume I own shares in every company I write about, and read accordingly. I don’t disclose individual positions or sizes.
Disclaimer. This is analysis for discussion, not investment advice, and not a recommendation to buy or sell anything. I’m not your adviser. The figures come from primary filings but I make mistakes, so check them yourself before you act. Micro caps carry real liquidity and volatility risk, and you can lose your money.