Description

Short Tucows

Overhyped Ting Internet division paired with deterioriating fundamentals in the mobile and domain businesses reverts as metrics continue to lag implied potential

Investment Summary

Tucows (TCX), comprised of two stagnant internet businesses and a nascent fiber broadband business, should decline materially as the market comes to understand the fiber business cannot deliver on the FCF that the market is pricing in. The stock has been driven almost entirely by the narrative that their CEO has propogated about their Fiber business: that Fiber delivers $1,000 of gross profit on $2,500 to $3,000 of capital investment with the ability to deploy nearly endless capital at that return. We believe that Tucows can neither deliver the stated economics nor scale the business to the size that its valuation demands, while its other businesses are poised to stagnate or decline:

• Fiber unit economics: management anchors investors on $1,000 gross profit on $2,500 to $3,000 of capital investment at maturity. We believe that both the model assumptions and the framing are incorrect, and that this business will struggle to earn its cost of capital

• Assumptions: management relies on unrealistic beliefs for FCF potential, ultimate penetration, and particularly capital requirements

• Framing: management emphasizes gross profit relative to capital investment at maturity, while we believe the relevant frame is the total IRR over the life of the project, including the impact of SG&A, maintenance capital, and the substantial ramp-up time where that capital earns minimal or negative returns

• Fiber unit scale: while the fiber market is theoretically huge, we believe several constraints limit growth

• Opportunity: Tucows has made it clear they only want certain types of communities and certain types of deals, and they have already turned down seemingly prime opportunities

• Management capacity: this is management’s view of the primary limiter, and we believe that they will indeed struggle to grow the business to scale over time without adding meaningful management/overhead expense

• Capital: Tucows has stated it does not intend to raise capital to fund this opportunity, limiting growth to its internally generated cash, or ~100,000 of home passed over the next 5 years

• Ting Mobile: Ting Mobile, Tucows’ MVNO, is clearly a source of cash but declining gross adds puts the overall trajectory of the business on a downward slope as competition heats up

• Domains: like Ting Mobile, Domains is a stagnant business that has historically lost share, offset by acquisition and pricing opportunities that have been close to fully utilized at this point

By applying more realistic assumptions and valuation multiples that better reflect the earnings trajectory of the respective divisions, we get to a ~$37 value, representing a 43% downside from the current price.

Business Description

Divisions, Geographies & Exposures

Tucows is an internet conglomerate with three separate business divisions (exhibit 1):

1. Ting Internet: internet service provider that provides households with gigabit internet speeds at $89/month for consumers and $139/month for businesses in selected towns. In most cases, TCX works with municipalities and leverages their fiber infrastructure

2. Ting Mobile: MVNO that works with T-Mobile and Sprint. Operates a pay-per-use post-paid pricing model

3. Domains: domain registrar business which primarily services wholesalers like web hosting companies

Exhibit 1

While Ting Internet (included in the other network service) is currently a small piece of the total business, the market values this segment as a large chunk of TCX’s value.

Thesis Support

Ting Fiber unit economics

Model Assumptions

Management has repeatedly indicated to investors that TCX can achieve $1,000 in gross profit and $2,500 to $3,000 in capital investment per connected home. We believe these targets are both misleading and unachievable.

It is misleading as it fails to account for expenses below the gross profit line. It also depicts the business at maturity and doesn’t recognize that there will be many years where the capital generates minimal to negative returns as the business scales. It has also failed to highlight that it doesn’t own the fiber infrastructure in many of its existing and future regions, and therefore will not capture the economics.

More importantly, we believe these targets are unacheviable as it uses aggressive assumptions around the capital investment necessary to pass a home and the penetration rate at maturity.

Adjusting these assumptions to more achievable levels suggest that this business will struggle to earn its cost of capital, aligning with intuition and historical precedent. If well-funded companies, such as Google, were unable to make overbuilding work, what makes TCX, a domain registration and MVNO business, any different?

FCF per connection

Management has anchored investors on $1,000 of gross profit per connected home which is based on $89 per month per residential home of ARPU and minimal COGS. We believe this is right. However, it is an inaccurate depiction of TCX’s earnings power as it fails to recognize that substantial cost is necessary to service customers, to acquire customers, and to maintain the fiber infrastructure.

Profitability falls to ~$600 per connected home once these expenses are accounted for (exhibit 2). These expenses were estimated by benchmarking against peers and looking at TCX’s existing network access cost (exhibit 3), and further comparing against the profitability of its closest comp, Allo Communications, when it was on a slower growth trajectory (exhibit 4).

Lastly, management’s framing doesn’t capture any incremental capex needed to maintain the infrastructure as well as the time value of money which meaningfully alter the economics as it will take the company many years to reach maturity.

Exhibit 2

Exhibit 3

Exhibit 4

Capital investment per connected home

Management believe that it would cost $2,500 to $3,000 of capex to connect each home which is based on $1,100 to $1,400 of capex to each passed home, 50% penetration rate, and ~$200 to $400 of capex for each home connection.

We believe this range is unachievable as the capex per passed home is too low while the mature penetration rate is too high.

Capital investment per home passed

Management believe that it costs $1,100 to $1,400 of capex for each passed home, while we believe it is going to be closer to $1,600 to $2,300, with reality closer to the high end. We triangulated to this range by looking at mature cable companies, another FTTH overbuilder, TCX’s own reported data, and conversations with installers.

Mature cable companies capex per home passed suggests ~$1,640 of capex per home passed (exhibit 5). Closest comparable, Allo Communications, a small FTTH overbuilder in Nebraska, records $2,281 of capex per home passed (exhibit 6). TCX’s own reported data suggests $2,271 of capex per home passed (exhibit 7).

Our conversation with installers suggest that this range is reasonable, and TCX likely operates at the higher end of the range. Fluctuation in capex per home passed is dependent on two factors - household density and build type. It is cheaper to pull fiber to high density areas like to urban multi-family homes as one can pass more home, and it is also cheaper to do aerial builds than buried build because one doesn’t need to dig up the ground. This is likely the reason why Charter and Comcast have a lower per passing cost than Allo Communication and TCX as they have a mix of aerial and burial and connect to a mix of multi-family and single-family homes. Allo Communication and TCX, on the other hand, almost exclusively do burial fiber to low density suburban single family homes, which raise the cost to pass.

“It typically cost $500-$800 per home passed for aerial build in suburban areas. Home density is the biggest driver that swings this metric…underground (buried) build can be 3x the cost of aerial”

 - EVP Network Operations of a large telecom provider

Lastly, while it is true that TCX is often leveraging on some of the town’s fiber infrastructure and paying the town a network access fee for it, our checks tell us that the capex benefit is immaterial to the total required capital investment as evident by TCX’s capex per home passed of $2,271.

Exhibit 5

Exhibit 6

Exhibit 7

Attach rate

We also believe it would be challenging for TCX to deliver on its 50% attach rate (or penetration rate) target. Instead, attach rate closer to 40% is more likely (although still a stretch). This is because very few people actually require gigabit internet and will therefore pick cheaper options by the incumbents. For another, many incumbents already offer gigabit internet (AT&T in Holly Springs, Xfinity in Westminster). Lastly, TCX doesn’t offer TV or voice which makes its product less attractive for many customers. As such, we believe that there is nothing that differentate TCX, and in fact suggests that TCX is disadvantaged in many situations, which makes its target unachievable.

We got to our projections by triangulating against what mature peers have done, a small sample set of overbuilders, and further compared it against the number of internet subscribers the company would have if management’s assumptions were true.

While it is true that on average mature peers typically operate near 50% attach rate (exhibit 8), it is because these players run largely uncontested in many of their regions. This is not the case with TCX as the company is entering into areas with many incumbents. For instance, Holly Spring have CenturyLink, AT&T, and Spectrum. As such, TCX’s attach rate will likely be south of mature peers. We believe that at best, TCX attach rate will approach Verizon FiOS levels (a level Verizon only achieved after 13 years of operations).

Exhibit 8

Despite this, management appears confident in its 50% target as it compared itself against the performance of a small sample set of overbuilders. For instance, through a survey (door-to-door survey in 5 neighbourhoods) Bernstein Research found that Google Fiber achieved 75% attach rate in certain Kansas City neighbourhoods. This view is reinforced as the company’s existing regions are tracking in-line with expectations if we perform a cohort analysis and linearly grow attach rate from 20% to 50% from year one to five.

We believe these two processes are flawed.

For the former, TCX assumptions is based on a small sample set. While it is true that Google Fiber has 75% penetration rate, it is only in a couple of neighbourhoods and might not be representative of the opportunity at scale. As TCX enter less attractive neighbourhoods, we expect less success.  

For the latter, while it is true that several regions are tracking in-line with expectations, this doesn’t account for the fact that TCX has selected the best neighbourhood first where it pre-signed many customers, and doesn’t capture the increased competition TCX will have as it battle against higher customer attrition as its customer base expands (same gross add but higher churn as customer base is larger). Hence, growth pace will quickly falter as TCX move into less attractive regions and growth base expands.

Framing

Putting all of these together tell us that TCX will struggle to achieve an IRR that meets its cost of capital (7.9% IRR, exhibit 9) which stands in stark constrast with the way management framed the opportunity (38.2% IRR, exhibit 10). To re-iterate, this is because management failed to include a bunch of relevant cost, overstated the ultimate penetration rate the company would achieve, and understated the capital investment necessary to build the fiber network.

It should also be noted that our base case IRR estimate of TCX value is likely still too generous as there is a good chance the company will achieve attach rate significantly south of 40%, will take much longer than expected to reach maturity (remember it took Verizon 13 years to reach 40% penetration with FiOS), and there is likely even more cost such as maintenance capex that has not been captured.

Furthermore, management has misled investors by failing to highlight that TCX doesn’t own the fiber infrastructure in many of its existing and future regions, and as such will not capture the value. The company does disclose this fact, but then emphasizes only the economic numbers for fully owned systems. For instance, the company doesn’t own the fiber in the city of Westminster and has to lease the fiber. It also will not own the fiber infrastructure in the city of Fullerton and will only provide provisioning and customer service. In the first case, TCX will have to pay the city a fee to access the fiber and in the second case TCX will likely be paid a fraction of the full economics

We believe the market will converge to our view over time as TCX put out evidences that conform to our view – poor earnings realization, higher than expected capex per home passed, and slowing subscriber growth.

Exhibit 9

Exhibit 10

Ting Fiber unit scale

Even if the unit economics are better than laid out above, we believe it will be difficult for TCX to expand its fiber operations to a level that adds meaningfully to its share value as the company has limited opportunity, management capacity, and available capital.

Opportunity

Our checks tell us that management is looking for areas with highly specific characteristics – places where there is demand for gigabit internet but limited gigabit internet offerings, sufficient population density to justify laying fiber, and the ability to work with towns and leverage on existing infrastructure. For instance, Holly Springs and Westminster are both fairly dense and affluent towns that have agreed to allow TCX to leverage on the town’s existing fiber infrastructure.

While this is great as TCX is taking a measured approach, this limits the opportunity set and growth pace as there are not many towns that fit these characteristics. This limits the ultimate size of the business and reduces the pace in which TCX can grow.

Capacity

Moreover, our checks tell us that TCX’s management already has its hands full with its current regions and there is limited bandwidth to accelerate growth. While adding capacity to scale its operations is possible, doing so would add to meaningfully to costs.

Capital

Lastly, growth is further limited by capital. Pulling fiber is expensive and management does not intend to raise capital. This limits the company’s growth to the amount of cash TCX can harvest from its declining mobile and domain division, lowering its present day value due to the effects of discounting (although this is counter-intutively a positive for the company if the IRR is ~7.9% as estimated, as it limits the company’s ability to destroy value).

To give some perspective on growth pace, even if TCX allocates all excess capital to its fiber business, it will only be able to build out ~100,000 incremental home passing over the next 5 years (exhibit 11). Moreover, this number assumes that the cash flow from the mobile and domain business will remain stable, when in reality it will decline. Hence, even if TCX does generate some value from the fiber business, it would been challenging for the company to grow into its valuation.

Exhibit 11

Yet, the market appears to believe that the fiber business is worth ~$372mn, approximately half the company’s current enterprise value, which even at management’s case assumes ~135,000 home attached, or ~103,000 incremental home passing (exhibit 12). Hence, even with management’s highly aggressive assumptions, it will take the company 5 years to grow into its current valuation, and, since those values are based on discounted values, the break-even number of passings will continue to go up over time.  

Exhibit 12

Ting Mobile

Ting Mobile trajectory

We believe that Ting Mobile is on a downwards trajectory as the business sees increased competition from other MVNOs that offer similar, if not better, products. This is simply a continuation of recent trends.

Business history

TCX launched Ting Mobile in 2012 and had been very successful with the business as it grew revenue to $89mn by 2018. Ting Mobile was able to achieve this by providing a unique value proposition to a niche and underserved market segment: low-use phone users. It allowed customers to customize plans to their needs. So a person that rarely called or texted can configure a phone plan at a much smaller cost than with the national carriers. For instance, the average Ting Mobile customer spends $30 per month per line, which is significantly cheaper than T-Mobile which charges $70 per month per line.

Competitive landscape

However, there was nothing inherent about the business that prevented its replication and Ting Mobile’s success attracted a flood of competition. New players such as US Mobile and Mint Mobile offer similar (if not better) packages at lower prices (exhibit 13).

Exhibit 13

*500 minutes talk, 1,000 texts, and 500MB of data

This increasingly competitive environment is reflected in TCX’s shrinking customer base over the past two years (exhibit 14), predominately from declining gross add against a stable churn (exhibit 15). We believe this trend will persist as players continue to capture share from TCX by offering a comparable, if not better, product.

Exhibit 14

Exhibit 15

*Excludes 22,000 acquired RingPlus customer in 1Q17 from gross add. Many RingPlus customers subsequently migrated in 2Q17.

Domains

Domain trajectory

Like the Ting Mobile business segment, we believe the domain business will continue to decline as it has in the past interspersed with small acquisitions to offset the trend.

Business history

The domain segment is TCX’s original business (prior to Ting Fiber and Mobile). It provides domain name registration services primarily through its OpenSRS and eNom brands. Its customers are primarily web hosting companies which typically bundle domain names with higher ARPU services like web hosting. This service accounts for a small proportion of a web hosting company’s total cost, but is often necessary to acquire customers.

Competitive landscape

However, the issue is that domain registration is a commodity and there is nothing that really differentiate one company from the next, making price the swing factor in the purchasing decision.

Comparison to peers tell us that TCX is not the price leader. In fact, it is priced at a sizable premium to other players such as NameCheap and GoDaddy (exhibit 16). This price premium came partly as a result of a series of price increases in FY17 and FY18 (exhibit 17). However, while these price increases temporary benefited revenue growth (exhibit 18), it also accelerated share loss (exhibit 19, 20).

TCX has been trying to mask this decline by acquiring domain registration companies such as Melbourne IT, eNom, and most recently Ascio Technologies (exhibit 21) and while TCX may continue to do this, we believe the opportunity to do so has shrunk as there are not many sizable domain registration companies left. The company also lacks to capital to grow via acquisition as the company has to deploy capital to its fiber initiative and is already fairly levered.

Hence, to stabilize the situation, TCX must lower its price. However, doing so will significantly impair its economics as this division has high fixed cost in the form of registration fee to Verisign. Hence, the domain business is in between a rock and a hard place where the company either loses share over time or impair its unit economics.

Exhibit 16

*first-year pricing

Exhibit 17

Exhibit 18b

Exhibit 19

Exhibit 20

Exhibit 21

Valuation

In light of the above, we believe TCX as a business that is dumping its cash from its diminishing mobile and domain divisions into its value destructive fiber business. Given such a dynamic, we believe TCX is worth $37 per share at best, and even less as management continues to funnel capital into the fiber business. This value was estimated through a SOTP approach to account for the divisions different dynamics and capital intensity.

Mobile and domain value

Ting Mobile and the domain business were valued at 10x P/E, or approximately $416 million (exhibit 22). We believe that a 10x earnings multiple is appropriate, if not generous, given that the divisions are quickly melting ice cubes.

Exhibit 22

Ting Fiber value

Valuing the fiber business, on the other hand, is trickier as the more TCX invest in the divison, the less it is worth. It could potentially have a negative value. For the sake of conservatism, this business is valued at $61mn (exhibit 23) as it currently stands and doesn’t not assume incremental investment (and therefore value destruction).

Another way that we can value TCX’s Ting Fiber installed base of home attached is by estimating its real estate value. We can do that by working backwards and solving to a capex per passing assuming a 10% discount rate. Doing this suggests that each home passed is worth ~$1,984, which tells us the Ting Fiber installed base is worth $60mn (exhibit 24), aligning with our previous triangulation.

As TCX continues to dump capital into the fiber business, the fiber division value will decline. At its current pace, we believe TCX will destroy ~$3.5mn of value, or ~$0.32/share annually (exhibit 25).

Exhibit 23

Exhibit 24

Exhibit 25

Total value

All of this suggests that TCX is worth ~$37 per share, and potentially less as the company invest in the value destructive Ting Fiber business (exhibit 26).

Exhibit 26

Catalyst

Earnings releases and metrics reflecting Fiber business failing to achieve sufficient economics to support valuation