Weekly Commentary 13/09/21: I contain multiples

The FTSE 100 was down -1.4% this week to 7,048. The Nasdaq and S&P500 also fell, but less than -1% each. The FTSE China 50 was up +1% last week, the Mumbai and Russian stock exchanges remain the best performing YTD, both up +25% since the start of the year.

Oxford Nanopore, the maker of DNA sequencing devices that track variants of Covid, announced plans to list on the London Stock Exchange, at a mooted valuation of £4bn. This is an IP Group success story, which I wrote about a couple of weeks ago. IPO owns a 14.5% stake valued at £359m on the 30th of June (implying a value of £2.5bn for the entire company). If that £4bn IPO price is achieved, it would imply a valuation of around 33x historic revenue.

When I started in The City at CSFB, analysts tended to see valuation multiples as a “short cut” and instead spent much of their time building very sophisticated Discounted Cash Flow (DCF) valuation models, with many inputs and different assumptions to generate a target price. These large, intricate models were used to create impressive complexity and justify seemingly absurd valuation multiples. I once sat next to a young insurance analyst who spent several weeks building an elaborate valuation model, that after much huffing and puffing, churned out a target price within a couple of percentage points of the current market price of the insurance company he was trying to value. My work colleague was rather proud of his achievement, until I pointed out that it was the equivalent of building a sophisticated computer weather model, designed to tell you if it was currently raining outside, when the simple way of telling if it was raining was to just look out of the window.* If you’re building a fancy model, it needs to generate a target price at least 30% above the market price to pique the interest of fund managers.

Then following the TMT bubble bursting post September 11 and later the financial crisis in 2007-8 these fancy models lost credibility. Now we don’t seem to need fancy models to justify high valuations. For instance, on AIM there are currently 130 companies (of 853) trading on more than 10x sales, and 716 companies (of 4560) on more than 10x sales on Nasdaq.

I came across an interesting essay on valuation multiples from a couple of chaps at Andreessen Horowitz, the venture capital firm. When VC’s put a value on an early-stage firm in a funding round, it’s very hard for founders to understand why their company is “only” worth 10x revenue, but another similar start up is worth 20x, 50x or a 100x revenue multiple. The AH writers explain that they don’t spend much time building intricate DCFs models, but trying to imagine what the company might look like in 5 years time and what multiple investors would be willing to pay then.

They use the example of Salesforce to illustrate the point. When Salesforce went public in 2004 as a new kind of Customer Relationship Management (CRM) software, its S-1 filing (the IPO prospectus that US companies file as an Admission Document) showed revenue of $96m and indicated the entire CRM applications market was worth $7bn. But since then, the company has expanded its scope, and FY to Jan 2021 reported $22bn of revenue (and is still forecast to grow revenue at +18% this year). Andreessen Horowitz point out that if an investor could have identified Salesforce’s ability to maintain growth, invested in its 2004 IPO, and then held on through to today, they could have made ~70x returns. Not too shabby!

In fact, you would expect as a company increases in size, that it becomes harder to maintain the same pace of growth and the valuation multiples contract to be comparable with mature companies. Yet SharePad shows that Salesforce even now trades on 11.5x price to sales, well above the historic average of the last decade.

This week I look at four companies, all with net cash on their balance sheets, but trading on vastly different valuation multiples. Counter intuitively the least profitable (Bango and Boku) are trading on the highest multiples (>13x sales), whereas the most profitable (Somero and Sylvania) are much more attractively valued (4.8x and 1.8x respectively).

The comparison on the next page implies that either i) future revenue growth will be much higher for the software companies ii) their returns will improve substantially as little incremental capital is needed to fund growth, or iii) SLP and SOM are cyclical industries, reporting close to peak revenue and profitability and future results are likely to disappoint.

All of those things may be true, but it’s worth asking: what are these price to turnover multiples already pricing in? I’ve used SharePad’s “compare” table and exported as a CSV file to make the point clearer.

My instinct is Boku and Bango trade on higher valuation multiples, not because anyone has built an elaborate DCF model, but because similar peer group companies trade on the same high multiples. Yet mobile payments, like adtech, is likely to be a sector with a few winners earning the bulk of returns, and many “also rans”. It makes no sense for every payments or adtech company to trade on the same high multiples that assume success. Even though loss-making Amazon in the late 1990s did go on to justify its valuation, you didn’t want to pay a high price for Webvan, Boo.com and Pets.com in the late 1990s bubble.

Boku and Bango H1 June results

These two mobile payments platforms are not direct competitors, but they operate in a similar space (mobile payments and direct carrier billing), so I thought I’d cover them together. Both are trading on c. 13.7x sales, and over 35 EV/EBITDA multiples but struggle for profitability.

Bango helps app developers and online stores earn money from payments data. The company also lets app developers and merchants focus their marketing spend on users who have previously paid for similar content, which increases the marketing “bang for their buck”. Bango reported H1 revenues +49% to £7m, adjusted EBITDA +83% but still no profits (loss before tax of £450k). They had net cash of £7m at the end of June.

Boku, which is 4x larger than Bango and counts Microsoft, Apple, Amazon, Google, Netflix and Spotify as customers, allows users to pay for digital goods with their phone, with Boku’s Mobile First Payments Network. The Chief Exec, Jon Prideaux, uses the example of Thailand to illustrate what Boku does. If Europeans or Americans want to travel to Asia and buy anything in Bangkok, they just take a visa card. But if European or American companies want to sell anything to Thais, accepting Visa is not enough, they need PromptPay, Rabbit LinePay and TrueMoney, because that’s how people in Bangkok pay for digital services. These are hard problems to solve, there are no standards to follow, tax, commercial and regulatory issues are a headache, so Boku solves this problem for the likes of Apple and Spotify etc.

Boku reported revenue +38% to $34m (of which organic revenue growth was +21%), adjusted EBITDA +61%, and PBT $1.9m (a 22x increase vs just $87K reported last year). Boku had $48m of cash at the end of June, but I think that figure should be reduced by $11m of debt taken on to fund an acquisition last year (Fortumo) so that net cash is actually $37m.

Valuation Both of these companies look expensive, even assuming that they grow revenues on relatively fixed costs, the 2023F forecast suggest 56x PER for Boku and 45x PER for Bango.

Opinion It’s good to see both companies making progress with their “platform” strategies (perhaps Boku more so than Bango). But they don’t fit my process, I prefer to see high returns on capital as evidence that the strategy is working, rather than just revenue and adjusted EBITDA growth. If the valuation was more conservative (say 5x sales) I’d see the risk / reward profile as favourable, but at 13x sales that doesn’t leave any room for either stock to disappoint. I’m also rather nervous that there are some big players in this area, Stripe but also Tencent and Alipay who might develop a competitive product. It’s tempting to suggest the economics of mobile payments are more favourable than platinum or concrete screed machines (see companies covered below), but for precisely that reason investors need to have a high level of conviction that the competitive “moats” justify paying >10x sales.

Sylvania Platinum FY to June 2021

Sylvania, the South African Platinum Group Metals (PGM) company, released FY results to June. It’s not really a mining company, the core business is the retreatment of chrome tailings from “host” mines to produce PGMs, which means that SLP enjoys a cost advantage versus companies that dig platinum and the other metals out of the ground.

SLP had already announced headline figures in a Q4 update at the end of July, saying revenue fell -35% to $48.4m vs Q3, as their average 4E gross basket price of PGMs fell -11% to $4,059/ounce. Last week’s FY figures give the full financial statements. On a FY June 2021 vs FY June 2020 basis, revenue was +80% to $188m, net revenue which includes a “sales adjustment” for ounces already delivered in the prior year but invoiced in the reporting period, increased +79% to $206m. Net profit (which includes an increased mining royalty tax) was up +143% to $100m.

Cashflow Adjusted EBITDA was up +108% to $145m. Notably a reconciliation of $143m FY PBT to FY cash generated from operating activities of $113m is tucked away at the back in Note 24. The bulk of the difference between the two figures is a negative $34m movement in working capital, due to an increase in trade and other receivables (ie money that is owed to SLP.) At the end of June, Sylvania had $69m of trade receivables on their balance sheet, of which $47m was subject to provisional pricing. That is, the platinum has been delivered, but not yet paid for and is exposed to falling commodity prices.

Though it’s a small figure in the context of FY costs of $54m, I noticed that the depreciation charge has halved to $2.8m (increasing profits) as management increased the useful life of equipment assumption. Presumably if they hadn’t made this accounting change, the +17% FY growth in cost of sales would have increased by around 4% points higher than actually reported.

Chrome host mines Much attention for the SLP investment case has focussed on the Rhodium price, which spiked higher earlier in the year to $25,000 an ounce. This meant that for SLP the average PGM gross basket price during their financial year was $3,690/ounce; a +83% increase on the previous year. Rhodium is currently $13,260 per ounce, and down -30% over the last 3 months, which implies that the PGM basket price will be down sharply in the coming year.

However rather than Rhodium, the Chairman’s statement suggests SLP’s reliance on the chrome mining market is more important for the long term outlook. Chrome mines have struggled in recent years and this means that host mines are producing less tailings, and SLP unit costs have risen, as the group has to process +17% more feed tons to deliver the same numbers of PGM ounces as the end product (production was up just +1% to 70,043 ounces).

However, in their investor presentation, management were still able to put up the Nedbank cost curve chart showing that SLP is still in the best quartile of low cost producers (red arrow in chart below) and reiterate their 70k ounces target for this year and next.

Opinion The last year has had some helpful tailwinds (Rhodium price) offset by headwinds (power disruptions, Run of Mine costs rising, the mining royalty tax and of course Covid.) Management can’t control the price of PGMs, but they do seem to be operating a strong business and focussing on what they can control. It does seem to me that if the chrome miners are really struggling with a downturn, this could be a risk for the Sylvania investment case. But at the current valuation, it appears that much of that downside is already priced in.

Somero H1 results

This US headquartered concrete screed company reported their highest ever H1 revenue, +82% to $64m and profits trebling to $23.5m. The previous highest H1 revenue was June 2018 of $45m, so despite H1 2020 being an easy comparative, these are still very strong results. The group ended June with $32.8m of net cash on their balance sheet.

Activity has been driven by non residential construction activity in the US, particularly demand for new warehouses (with quality concrete floors) to keep pace with the shift to e-commerce and online retailing. There was also some “catch up” effect, versus reduced activity last year. The only slight disappointment was China, which is less than 3% of group revenue, where the business went backwards. A couple of years ago China was supposed to be a big part of the growth story, Somero has low market penetration and the country is 50% of global cement production. Yet management now say that there’s simply a lack of demand for quality concrete floors. I suspect that there’s not much Somero management can do, if Chinese building owners are not prepared to pay up for the quality and support, it may take a while for the negative effects of local “bodge it and scarper” builders to be felt.

Outlook That aside, the future looks positive. Management are signalling confidence by adding 50,000 sq feet of capacity (a 50% increase) at a cost of c. $10m. This will raise capacity from $130m revenues to $175m and is expected to be ready mid 2022. The board has raised guidance and now expects FY 2021F revenues will approximate $120m (previously $110m). No PBT guidance, but adjusted EBITDA should be $42m (previous guidance $35m). They now expect year-end 2021 net cash will be c. $36m (previously $33m) which includes the 2021 spending on the facility expansion project. Given H1 was $64m in revenue (ie implying a decline to $56m in H2) that looks like it could be conservative, but management mention some of that record H1 figure was driven by backlogs from the previous year, plus there’s a risk of Covid resurgence and supply chain disruption. I think that this is the third increase in guidance this year, the most recent upward revision coming in July.

Broker forecasts FinnCap, their broker, has raised their FY 2022F revenue forecast by 8.5% to $128m, but no 2023F, I suspect because management are reluctant to commit to a figure, knowing how cyclical their markets can be. So 2022F EPS of c57.5 implies a PER of 13x (1.4x USD/GBP forex rate), which seems good value for a company with gross margin of 57%, forecast 2022F ROCE of 55%. FinnCap is forecasting 41c of dividends in 2021F, putting the shares on a generous yield of 5.6%.

Opinion I’m normally dismissive of US listed companies coming to AIM (after all if Manchester United, dozens of SPACs and a New Jersey Deli with sales of less than $37k can list in New York, most viable US companies should find it easy to raise money on their home stock exchanges and don’t need to come to AIM.)

Somero, after a shaky start on AIM has been a success story and I’m glad I bought the shares in 2015 (though once again I was too cautious on the position sizing). Over the last 10 years the shares are up +5070%, which is better than Games Workshop +2570% and Impax AM +2250% over the same time period.

The one cloud on the horizon is the age of management. Lawrence Horsch, non-Exec Chairman is 87 years old, Jack Cooney the Chief Exec is 74 years old. I rather like an experienced management team in a cyclical industry (Adam Applegarth’s was in his 30’s when he became Chief Exec of Northern Rock) but I’d be a little nervous if management decide to stand down, as we’ve seen a few times (eg Accesso) that can be an indicator that we’re close to the peak of the cycle.

Bruce Packard


*He now works in investor relations.

The author owns shares in Sylvania Platinum and Somero

Leave a Reply

Your email address will not be published. Required fields are marked *