|Shares Out. (in M):||41||P/E||-13.5||-20.3|
|Market Cap (in $M):||442||P/FCF||0||0|
|Net Debt (in $M):||4||EBIT||-27||-20|
|Borrow Cost:||General Collateral|
I wrote up Boingo as a short in August 2015, just after the company released its quarterly earnings results and the stock had spiked to $10, $1 above its previous recent three-year high. I said:
My bullish scenario for the business, if everything goes as management hopes, yields an $8 DCF valuation for the stock, 20% downside. I think it more likely that something goes wrong and fair value is closer to $5, 50% downside. This downside could come quickly if something bad happens to the business, or, if everything goes well for the business, it could come through years of alpha as the stock stagnates while the S&P 500 pays dividends and marches higher. The likely result is somewhere in the middle: The stock will probably present good trading opportunities from the short side, bouncing around such that you can cover lower and re-short higher.
That prediction has proven correct. The stock promptly fell 25% in the three weeks after I posted, versus 4% for the S&P 500, for 21% alpha, and I covered the position both in real life and in VIC-space. It continued falling to $5.50 in the January/February market dump before recovering sharply along with all the other dodgy stocks. Today, two trading days after the company’s latest quarterly earnings release, the stock has spiked to $11, $1 above its high achieved just before my previous write-up. I have re-shorted and invite you to join me. I have enough to say about the last 15 months’ developments to merit another write-up rather than just a comment-board post. The thesis remains valuation plus a catalyst that is not precisely knowable yet is reasonably likely among a host of possible candidates.
See my earlier write-up for a detailed description of Boingo’s various business lines and how they were likely to develop. Almost everything has played out as predicted. The net effect of all the news is to make me slightly more negative about the business while the average investor/trader is probably more positive. To summarize the developments: (1) The overall financial picture is worse than I thought, margins are not improving and now appear less likely to ever improve the levels required for my $8 bull case; (2) growth is about to fall to near-zero in Boingo’s good business (military) while it is accelerating in the bad business (DAS); (3) reported revenue growth will probably turn negative in the medium-term because so much of it is up-front build-out payments from carriers.
Now here are the detailed updates, starting with each business line:
Military (25% of total 3Q revenues) is probably Boingo’s best business by far because it has an effective monopoly with contractually-fixed prices and a bargaining partner (the Department of Defense) who is unlikely to turn the screws when it’s time for contract renewal. However, as predicted in 2015, this business is about to fall from triple-digit growth to near-zero growth. Boingo has been rapidly installing networks at new military bases, increasing the “covered bed” count, increasing the subscriber count even faster as subscriber penetration of covered beds increased, and therefore growing revenues by 341% in 2015 and a still-high 91% in 3Q16. Management always said that they would hit full build-out of 300,000 beds by year-end 2016, and as of 3Q they are on-track at 286,000 beds. Growth in penetration of covered beds has slowed to one percentage point per quarter, and that sequential growth number is more likely to go down than up. So sequential growth will be near-zero starting in 1Q17.
Distributed Antenna Systems (DAS) (39% of total), which is Boingo’s worst business in terms of return-on-capital, enjoyed a sudden growth explosion in 3Q. The company added 500-800 DAS nodes per quarter in 2015, 1,600 in 1Q16, and 1,000 in 2Q16. During 3Q it added eight new venues (e.g. stadiums) and therefore 4,200 new nodes, which drove the YoY growth in nodes from 42% to 71% and DAS revenue growth from 15% to 22%. Management said 4Q will be another big quarter for added venues and nodes.
Retail wifi (15% of total) continued dying but has, at least for now, stabilized for three quarters at around 185,000 subscribers and $6.6m in revenue. For now, this development is a genuine positive. It remains to be seen whether this stabilization lasts. I doubt it. The business is still showing 15% YoY revenue declines and will not lap them to get to 0% until mid-2017.
Wholesale wifi (14% of total) stalled out from 45% revenue growth 2015 to -6% in 2Q16 and 7% in 3Q16 because Boingo lost a big customer in 2Q. Growth will probably pick up again as the lap the customer defection.
Advertising & other revenue (5% of total) is pretty much dead by now, falling 61% in 3Q.
The aggregate of business-line-level developments have not been kind to the key overall numbers and won't be in the future. Last year when I calculated that WIFI’s stock was worth only $8 in a bullish scenario, that scenario included EBIT margins rising from their then-current -15% to +13%, above previous all-time highs. If margins were going to rise anywhere near there, they should have started doing so already in 2016 because the business mix has shifted towards the better military business (up from 14% to 25% of total revenues), plus margins are presumably improving in the military business itself as penetration per bed increases (up from 26% to 30%). Yet company EBIT margins have been down 1-2% YoY each quarter in 2016. That development makes it less likely that future margins will rise anywhere near my earlier bullish case. With the military business about to stall out, where is the margin growth supposed to come from? Obviously the rapid recent DAS buildout has dragged down margins and DAS margins will increase as those networks fill up, but that future increase does not look like nearly enough to get total margins into double digits.
Free cash flow is better but still not good. Roughly half of WIFI’s doubling off the bottom since February is due to the general rise since then in all dodgy stocks, and the other half is due to management’s promise throughout the year that the business would have “positive free cash flow” in 2H16, which is better than I had modeled last year. The final stock price spike up following the 3Q release, despite management maintaining its revenue and profit guidance, was probably because they achieved that goal; operating cash flow less capex through 9M16 was $1.5m.
To management’s credit, positive FCF does mean that the business won’t burn through the last of its cash and need to raise capital in the near future. However, hitting the promised target does not suggest that there is much positive free cash flow in the company's future. To start with, management is playing the usual game of profit-challenged, cash-challenged businesses by increasing stock compensation much faster than the business is growing – in this case, by 50% YoY vs. 15% revenue growth. Stock comp has totaled $10m YTD. Thus free cash flow for valuation purposes, as opposed to how-soon-until-a-liquidity-crunch purposes, was -$8m.
More importantly, $64m of their YTD cash flow is an increase in deferred revenue resulting from up-front cash payments by carriers to fund their capex for the DAS network build-outs. Of Boingo’s $92m total capex, $66.4m has been funded by these up-front payments by carriers. Typically when a telecom company that is “spending for growth” hits FCF-breakeven, the path forward is substantial FCF growth as capex recedes and the network gets filled. That will not be true for Boingo. When its capex recedes, so will the build-out payments from carriers. It will be left with only the access fee payments, which financials to date suggest will be negative-margin forever.
As a final side-note, the build-out payments become deferred revenue and get recognized as revenue over the estimated life of each carrier contract. This amortized build-out revenue is currently 2/3 of all DAS revenue and 23% of total company revenue. When Boingo stops building new DAS nodes at such a breakneck pace, its reported build-out revenue is going to turn negative fairly quickly and, given the lack of other growth drivers now that military revenue is about to stall out, total company revenue may turn negative again. That won’t matter to how I view the business’s economics, but it will look bad optically to – I’m so glad I can borrow this phrase from cfavenger – “the stocktwits crowd” and anyone else who has not been paying close attention.
Revenue growth turns negative.
Stocks of money-losing businesses with no net cash take another dive many times as large as the S&P 500's (as in August 2015 and January 2016).
|Subject||Couple more Qs|
|Entry||11/09/2016 03:44 PM|
As you know I've always been suspicious of this company and think their "EBITDA" numbers are bullshit- their EBITDA mainly comes from recognizing the deferred revenue in DAS buildouts, the costs against which are reimbursed capex which go into D&A. So all they have to do is take on more zero margin reimbursed buildouts and they will increase EBITDA without making any money. However, I did see a long pitch from ValueXVail at http://valuexvail.com/presentations/#mg and while it seemed exaggerated it didn't seem completely crazy. At least these particular investors are not falling for the fake DAS buildout EBITDA, rather they seem to believe 2 things: (1) DAS maintenance revenue is high margin and should be accorded a high multiple similar to a tower company and (2) the company has a long runway in wholesale WiFi offload sold to the wirelss carriers. So my questions are:
1. Why do you think DAS maintenance revenue is negative or low margin? Seems like there should not be signifiant ongoing costs.
2. What is going on with WiFi offload? I know the ramp has been slower than expected but they did sign Sprint and one other Tier 1 carrer. What is your outlook for this?
|Subject||Re: The mystery of the cash flow statement - Specialk? Bueller? Anyone?|
|Entry||03/07/2017 07:05 PM|
just checked 2015. They book capex into ap and accrueds to start and they when they pay them they reclass to capex.
So when you buy it it debits ppe and credits AP and accrueds. Then it gets reclsssed subsequently typically when it's actually put in use.
Thats why you're assets work for the most part and your liabilities don't.
See the bottom of the 2015 cf statement.
Its not totally unusual to do this. I saw it at EY a fairly regularly. Fcf is the same net net.
|Subject||Re: Re: The mystery of the cash flow statement - Specialk? Bueller? Anyone?|
|Entry||03/07/2017 07:07 PM|
Here's the note from the 2015 cf statement.
|Subject||Re: The concise structural argument against DAS|
|Entry||03/10/2017 09:38 AM|
I don't think your comments on DAS here are even close to accurate. Let's go over them:
The management team like to say that DAS produces infinite IRRs (since they put up no capital), I don't think this is accurate since they have development costs, but the IRRs will be very high. Further, these same venues will be the next targets for wi-fi offload contracts. So far the wi-fi offload has been focused on 30 airports, but longer-term this will expand. We have spoken with several people who sell DAS contracts on venues that they have built out. The cash flow multiples on these deals at about 20x cash flows once installed.
|Entry||03/10/2017 04:16 PM|
This discussion is interesting, but I think it has gotten a little conceptual. Other than an "if you arbitrarily reduced their opex the stock would be cheap" I just haven't been able to find a compelling argument that this is cheap or even would be cheap if they kpet executing. The FCF is nowhere near the $31M that ATM stated, you simply took their EBITDA forecast (which includes DAS build-out reimbursed capex revenue at 100% EBITDA margin) and subtracted their non-reimbursed capex, when they are obviously also spending the capex they are being reimbursed for, which is why their FCF was only $8M for the whole year despite enormous increases in deferred revenue and why analysts only expect about the same this year.
I think the right way to look at this is to take their EBITDA ($40.6M in 2016) and subtract the portion that is reimbursed capex with no margin ($37.9M of DAS build-out revenue x 80% = $30.2M) and you get only about $10.4M of real-deal EBITDA. If you want to be more charitable to them you probably want to take DAS bookings less associated capex, which in this case (assuming the change in DR was all due to DAS) you have $37.9M of build-out revenue plus $71M change in deferred revenue less 80% or $87M in reimbursed capex or $22M in positive cash flow from DAS build-outs vs. the $37.9M claimed. If you subtract this difference ($16.1M) from their $40.6M in claimed EBITDA you get $24.5M in real cash EBITDA and this foots decently well with their actual FCF of $8M as according to this analysis there would be $107M headline capex less $87M of reimbursed capex or $20M in unreimbursed capex. The difference between $3.5M of calcualted FCF vs. $8M actualy could easily be explained by other items.
18x trailing EBITDA and a 2% FCF yield strikes me as expensive for this company, although it is growing double digits in revenue (somewhat understated if you look at DAS bookings) and may have some runway with DAS, military and possible WiFi offload. I do have some trouble with the WiFi offload piece though, I have been hearing about this forever and despite the years passing, the supposed bandwidth crunch, signing of two Tier 1 carriers etc. wholesale WiFi is declining and the conference call downplayed it as a contributor to next year's growth.
What is this worth? DAS ongoning revenue is nice high margin recurring revenue, I estimate it could be $28M in 2017 at 70% margins (they have some maintenance and revenue share with venues), roughly $20M per year of CF. If you want to put a 5% cap rate on this I guess you could say it is worth $400M alone although you would be conveniently ignoring all the expenses they need to incur to get this stream. The other problem is the rest of the businesses are negative FCF combined. I guess you could argue that military is worth $200M and the rest of the businesses offset the corporate and development expenses, but that is probably optimistic.
I am more persuaded by Katana's argument but am not short because I don't see why they can't keep growing fake EBITDA and don't know what makest it crack.
|Subject||Re: Hard Numbers - Massive undervaluation|
|Entry||03/13/2017 12:59 PM|
I spent a fair bit of time over the weekend working on separating the Boingo business into an Op. Co. which runs network assets and a Dev. Co. which builds network assets and depreciates the DAS equipment. To be clear, this is not an arbitrary reduction of expenses - the expenses are all valid and producing a tremendous amount of value, however, considering them all in one bucket leads to confusion and a massive undervaluation. You will see here that our Op. Co. is worth somewhere in the $16 to $21 range per share currently based on 2017 numbers. Further, there is nearly no Wi-Fi offload baked into this analysis - our estimate of value for Wi-Fi offload is up to $20 per share when you land all four tier-one carriers. In summary, the stock is massively undervalued based on the current income streams and even more so once Wi-Fi offload kicks in.