Description

Investment Thesis (Long): Internap (ticker: INAP)

INAP (price: $11.54) has 100-300%+ appreciation potential over the next 12-24 months. The company is significantly under-earning after a history of poor management produced an inefficient asset base --evidenced by the 59% occupancy rate and 32% EBITDA margins compared to peers at 85% and 45%, respectively.   Driving the re-rating is a restructuring led by a turnaround CEO (with significant skin in the game) that should produce top line growth and narrow the 10-15 point EBITDA margin gap with peers. If successful, INAP’s 50% valuation discount to peers (8x vs. 16x EBITDA) should narrow significantly.

The situation became even more compelling after a poorly worded press release alluding to an equity offering drove the stock down a further 35%. While the language was corrected several weeks later, the damage to the stock price had already been done (just look at the chart to see the technicals). This provides an attractive entry point -- management has made meaningful progress on the restructuring initiatives and investors today are able to buy the stock at nearly the same price before it was significantly de-risked ~18 months ago.

Situation Overview

INAP is an infrastructure as a service (IaaS) company offering datacenter colocation, cloud, and network services across 56 centers in 21 metropolitan markets. The company has a history of being mismanaged and the soap-opera includes a revolving door in the CEO suite, a failed sale process, and a footprint that seems to make little sense (utilization is 59%). Prior management executed the “GAAC” strategy – grow at any cost. Where INAP found a customer, the company was willing to open up a datacenter and allocate capital to the buildout. The outcome was 50% utilization (prior to new management) and a balance sheet that was 6x levered. Additionally, senior management did not focus on the sales organization, and the drawn-out failure to sell the company in 2015-2016 led to significant personnel turnover, customer discouragement amid false promises, and significant churn. However, a key positive we found in our research was the quality of the datacenters. The company had historically been run by engineers so the specs are highly competitive and meet the standards of Fortune 100 companies.

In September 2016, long-time shareholder Mario Gabelli (25% shareholder) had enough and personally hired turnaround CEO Pete Aquino to restructure the company. Pete has a solid track record having executed turnarounds at both RCN and Primus Telecom (PTGI). Prior to accepting the position, Pete did his due diligence and negotiated his compensation package to heavily favor equity given the upside potential. In addition to the 667k shares he currently owns (3.4% ownership, of which 200k was purchased in the open market before he started his job), Aquino will receive the equivalent of a $9mm payout as the stock reaches thresholds between $20-40.

Based on the current 8x EBITDA multiple, the market is valuing INAP as a highly distressed asset. Similarly leveraged companies (CONE, DLR, and QTS) are valued at 15x EBITDA (10x for the equity and 5x for the debt) compared to INAP’s equity value of 2x. The key discrepancy is occupancy levels (20-30% points) driving INAP to under-earn by 10-15 points on EBITDA margin. The stock price implies the situation will likely deteriorate while we believe it will improve significantly.

Why Does the Situation Exist?

INAP has always traded at a significant discount to the group based on its inferior fundamentals – occupancy, margins, and growth. As a levered, small-cap tech company, with a history of prior mismanagement, it is difficult getting new investors to hear the “turnaround” story, and the stock’s minimal analyst coverage does not help. While there are several reasons for the stock’s recent 50% decline, we believe the primary factor was the January 28 announcement to acquire SingleHop, which alluded to a potential equity offering by saying “ultimately, INAP will look to optimize its capital structure with a blend of equity and debt securities...” The timing coincided with general skittishness in the credit markets and a blowout of LIBOR OIS spreads. With an ominous sign of imminent dilution, sellers began to exit and few buyers felt the need to step up before a deal. It took a full month for clarification to be issued in a press release which stated “the Company has no immediate plans to raise additional capital.” By that time, the stock had already been tainted.

Despite meaningful tangible progress to date on the restructuring, the stock has little to show for it (the stock was roughly $10.00 when Aquino unveiled his turnaround plan to investors, peaked at $22 and sits at $11.54 today). We are somewhat baffled by the divergence of the stock price and fundamentals. For example, during the stock’s recent decline the following occurred:

·         Achieved sequential revenue growth in 4Q17 after several quarters of declines. This occurred three quarters sooner than sell-side analysts expected.

·         Signed several new name brand customers, one that is speculated to be Twitter. Signing a high quality customer further dispels any sentiment or belief that INAP has inferior assets.

·         Repriced its debt, saving $5mm a year in interest expense.

·         Generated $92mm of EBITDA in 2017 compared to initial guidance of $84-87mm.

Key Investment Points

Our investment case rests on the following:

1.       Rightsize the asset base to improve occupancy and margins.

2.       Drive organic growth by investing in the salesforce and cross-selling with SingleHop.

3.       Guidance is very conservative and our estimates are 15% higher than consensus.

Investment Point #1: Rightsize the Footprint to Increase Margins

INAP is a self-help story with too many data centers holding idle capacity. As a high fixed cost business, incremental margins only become significant as occupancy gets over 60-70%. The process of rationalizing excess capacity is not easy given customer preferences for certain data center locations, long-term agreements, and the disruption of moving facilities. However, management has started the blocking/tackling and has already achieved some impressive results. In some regions, INAP has multiple data centers in the same city (and even next to each other) and neither has proper occupancy (indicative of prior mismanagement).

The economics of capacity rationalization cannot be understated. In 2017, revenues declined 6% but adjusted EBITDA margins expanded 540 basis points as capacity utilization (of the company’s controlled datacenters) increased from 50% to 59%.

To illustrate the margin opportunity we have outlined the economics of a generic retail data center. This is not a specific INAP facility but demonstrates the important relationship between occupancy and contribution margin.  

As you can see, breakeven (at the data center level and not including corporate costs) is around the 32% occupancy level. At the 50% occupancy level, contribution margins are sub 30%. Peers such as CONE and COR are around 90% utilized and have ~55% EBITDA margins (at the corporate level). We believe that INAP should be able to get utilization up at least 10% (to 70%) and at that level will have >40% margins.

To provide further granularity of the margin potential, we will use INAP’s Boston facilities as an example. INAP currently has two data centers located next to each other and has stated that it plans to close one of its facilities in 2018. The company does not disclose PnL on a facility basis so we need to make several assumptions. For purposes of this example, we assume 50 Innerbelt will be consolidated into 70 Innerbelt. We also assume that INAP is successful in relocating 55% of the tenants (on a revenue basis). As shown below, while company revenues will decline by $7mm, operating profit will increase by nearly $3mm and firm-wide margins will increase by 90 basis points.

INAP is closing two other facilities in 2018 so we expect even further profit/margin contribution.

Investment Point #2:  Organic Growth Should Accelerate

The datacenter industry has many strong investment characteristics including long duration and secular tailwinds that should support significant industry growth over time. Companies differentiate themselves based on the quality of the assets, the location of the assets, and the breadth of services offered. While there is a good amount of competition, industry EBITDA margins and organic growth has averaged around 55% and 20% over the last several years, further cementing the durability of the business models and secular tailwinds.

In light of this, INAP’s performance is even more abysmal than the chart below would suggest. We believe this has been a direct reflection of poor management, C-level turnover, and the failed sale process. There were 50 salespeople when Pete Aquino took over the job and he deemed 30 (a whopping 60%) unfit for the role.

While rightsizing the footprint was the necessary first step towards INAP’s turnaround, the company will need to grow organically if it is to ever meaningfully narrow the valuation with peers. This does not understate the success new management has experienced so far. In 2017, revenues declined 6% but EBITDA margins increased 530 basis points. While there is still improvement to be had with asset pruning, organic growth will ultimately determine whether INAP is able to generate 35% margins or 45% margins.

There are two key factors supporting our belief that organic growth should improve relatively soon:

1.       Salesforce realignment has been completed and should produce results. When Aquino took over, there were 50 people in the sales organization. In 9 months, he eliminated 30 and hired 40 people. The massive churn further reduced sales given the disruption. By the end of 3Q17, the sales organization rebuild was complete. However, it typically takes new personnel 6 months to get up to speed and produce results so we have yet to see returns on this investment (even though the full costs are reflected in the PnL). We believe productivity should improve in 1Q/2Q18.

2.       SingleHop acquisition. We believe one of the key points to buying SingleHop was to accelerate customer acquisition growth. SingleHop has 3,000 clients across 6 data centers and a strong reputation on the sales side. Importantly, 93% of customers are in hosted cloud. This provides management flexibility on where to locate their data and will further improve utilization as INAP rationalizes its asset base. Additionally, there should be ample revenue synergies between the customer bases. Anecdotally, both management teams have already seen inbound interest on cross selling products.

The company has stated that its guidance implies positive organic growth, which we estimate at around 3%. The company has seen a high degree of churn, partially owing to some of the dynamics of the pre-restructured salesforce. As churn levels reduce and the turnaround progresses, we believe organic growth should accelerate to 5%+. In the last conference call, management stipulated that they “see an improving organic run rate.”

Investment Point #3: We think 2018 estimates are far too conservative

Most investors who have been involved with Pete Aquino and his team at prior companies would characterize their style as “beat and raise.” In 2017, EBITDA guidance was raised twice and the company achieved the high end of the range. We believe this is an important topic given that some investors may have been uninspired by 2018 guidance of $105-115mm in EBITDA. Given that pro-forma INAP/SingleHop (assuming no growth) is $106mm, management is assuming $2.5mm in cost synergies and accounting changes add $7mm, we find the guidance to be very low.

There are some other puts/takes to consider:

1.       Facility closures: management intends to close three facilities in 2018 and while this will be a hit to revenues, management stated on the Q4 call that it should be a positive impact to EBITDA.

2.       Cost synergies: we believe there is likely $2-3mm of additional cost synergies from SingleHop above management’s guidance.

3.       Revenue synergies: not included in management’s guidance but commentary suggests they are already seeing revenue synergies from SingleHop.

At first, we were somewhat concerned by 2018 guidance (i.e. is management being conservative or are there more issues in the business we are unaware of?). Our concerns are mostly assuaged given our experience investing with Mr. Aquino and understanding his management style, the anecdotal data points we have heard from industry participants, and the new customer wins.

Valuation

We believe INAP has the ability to achieve 40%+ EBITDA margins and mid-single digit organic growth. While this alone would make it hard to justify a peer multiple (which is growing ~20% with 50%+ EBITDA margins), it would narrow the valuation gap. On the flip side, we do believe that INAP could attract a strategic buyer after it turns itself around, which would augur for a higher multiple. All that being said, INAP is a levered equity so any strong fundamental improvement will likely disproportionately affect the stock price. Every 1x turn in the EBITDA multiple adds $6 or 50% to the share price. At 10x EBITDA, the stock would be worth 145% more and still be valued at 40% discount.

INAP has traded pretty consistently between 7-10x EBITDA and this is largely reflective of its inferior fundamentals. If fundamentals improve, so should the multiple. We also believe the company would garner a much higher multiple if all of it or single/multiple datacenters were sold. Per RBC, most of the datacenter M&A recently has occurred in the mid-high teens EBITDA range. When Mr. Aquino sold PTGI’s data center assets to Rogers he garnered a 14x EBITDA multiple and he believes INAP has higher quality assets.

Where We Could Expect Pushback to our Thesis

1.       High leverage and poor cash flow. INAP undoubtedly has a high degree of leverage (6x vs. 3-4x for comps) and adding incremental floating rate debt (for the SingleHop acquisition) in a period of increased credit market concerns does not help. Furthermore, INAP fails to generate significant free cash flow. So how is this sustainable?

There are a couple of points to note regarding INAP’s debt. First, the entire increase in the 2017 debt load was due to an accounting change as INAP reclassified its operating leases to capital leases. This added $180mm to the company’s debt and also increased EBITDA by $9mm. Had this change not occurred, 2017 debt/EBITDA would have been 1.5 turns lower. Second, the entire peer group is going to have to adopt ASC 842 which will require all operating leases to be put on the balance sheet as debt. This will add about 1 turn of leverage to the peer group, which will make INAP look better on a relative basis. Third, general concerns over the credit quality of the firm should have been put to rest on March 26 when the company announced it was able to complete financing for its acquisition and reprice its debt from L+700 to L+575, saving $5mm a year in interest expense. 

Regarding free cash flow, results have been poor but this is reflective of the subscale operating level. If our fundamental thesis holds, we believe the company will generate ample cash. At a 40% EBITDA margin, we think INAP would generate $30mm of FCF or about $1.50/share. Additionally, the company has publicly stated it is looking to divest non-core assets. An asset sale would not only help to de-lever but would provide an accurate mark on the company’s assets and show a large discrepancy with the public equity valuation.

2.       Why Should the Valuation Multiple Re-rate when Rackspace (RAX) was sold for 6x EBITDA in mid-2016. Pundits often cite this data point in suggesting INAP does not deserve a higher multiple. This is especially true since there are some similarities between RAX and INAP. Most notably, RAX operated at subscale levels with capacity utilization and EBITDA margins of 51% and 38%, respectively, similar to INAP’s subscale levels of 59% and 32%, respectively. Additionally, RAX also tried to sell itself as revenue growth peaked (2013) only to see the process fail and investors question the business model subsequently. This is similar to INAP’s failed sale process in 2015, which left questions on the quality of the firm’s assets. The similarities mostly end there and the key differences should point out why INAP is not another RAX. 1) RAX was a pure play cloud provider focused on competing at the hyperscale level (against Amazon Web Services, Google, and Microsoft). INAP is more focused on the SMB and enterprise market as a niche Tier 3 provider. 2) RAX decided to sell itself at the peak of its problems. Margins and utilization were low and revenue growth was decelerating massively. This is predominantly why the purchase multiple was so low. It was going to take several years to transform the profitability of the business. Strategic players in this industry do not like turnaround acquisitions as it is too much of a distraction to growth (this is similar to why INAP failed to garner a buyer in 2015).  It is no wonder that a private equity firm (Apollo) ultimately acquired RAX so that they could turnaround the business in the private markets.

3.       Are Shareholders Going to Be Diluted? We think this accounts for the majority of the stock’s sell-off. So why did management include language of a potential equity deal in a press release if it was not a likely outcome? We do not have a definitive answer, however, putting the pieces of the puzzle together here is our best guess (this is our conjecture and has not been corroborated by management). Given the company’s leverage when new management took over we believe it was always in the company’s plan to issue equity at some point. Within months of taking the reins, CEO Aquino did an early $47mm raise to shore up the balance sheet and refinance the debt. It is our belief that after significant strides were made in repairing the business, which were reflected in the stock price, the company may again do an equity deal to further repair the balance sheet. The 2016 10K even states the possibility of “the issuance of debt or equity securities or other possible recapitalization transactions” even after the February 17 equity deal was completed. As the year progressed and progress was made the stock peaked at $21 in October. At this point, management should have known they were going to have several positive data points to report soon including the new customer awards (including one we believe is Twitter), sequential revenue growth, and Q4 EBITDA beat. They were also likely working on the SingleHop acquisition and debt refi in conjunction. With the stock up significantly and management expecting a further pop given positive news flow, it is likely the company was considering an equity deal in the low $20s in conjunction with the acquisition and refi. The stock drifted to $17 as the market and data center sector sold off in November to when the SingleHop acquisition was announced on January 28. Given that an equity raise was likely contemplated throughout this process, we can understand why the company’s lawyers would require the line item be included in the press release. We believe that is the primary driver sending the stock down from $17 to $11. The company has made it clear that they do not plan on issuing equity at this time by removing such language from subsequent press releases. The company also recently refinanced its debt and an equity raise was not necessary to achieve that. Nonetheless, we believe this muddied the story.

Other Risks

1.       Aquino fails. The biggest risk is that Aquino fails in his turnaround. He originally signed a three year employment agreement indicating to us that he believed it would take that amount of time for the turnaround to succeed (he is half way through it). We were mildly disappointed when the Board amended his agreement to automatically renew in 1-year increments. We were told that this was purely for succession planning purposes. If Aquino does fail, he would likely have a difficult time branding himself as a turnaround specialist for his next venture.

2.       Cost of capital. One of the ways datacenter companies compete is on cost of capital and INAP arguably has one of the highest. Since the company is not in expansion mode (i.e. it does not need to raise capital to build new facilities) this is mostly mitigated at this point.

Catalyst

Improvement in organic growth, margins, and occupancy.