Game Digital — The (BE)long game

In February 2018 GMD signed a collaboration agreement with SD to roll out both BELONG gaming arenas and/or GAME retail units in standalone and SD stores. SD paid £3.2m for 50% of the existing BELONG business, comprising 19 units within GMD stores.

The agreement and what it says

The agreement states that GMD and SD will cooperate to develop BELONG gaming arenas and GAME retail operations in the following formats:

Each project will be jointly agreed before being committed to. Revenues and operating costs from the applicable operations (including a management fee to GMD) will be split equally between the parties. Where BELONG and GAME retail units are established within SD retail stores, the operating team will be controlled entirely by GMD. Rent will be negotiated as normal for a concession.

Fit-out costs will be borne by GMD, financed by a £35m long-term unsecured capex loan facility extended by SD, at 2.5% over base and with drawdowns repayable over five years following the second anniversary of each opening. SD is also providing a £20m one-year extendable working capital facility at the same rate.

The agreement is formally a profit sharing agreement and is not a joint venture agreement.

Acceleration of a key strategy

BELONG arenas, where customers experience the latest hardware and game releases, are key to GMD’s strategy of widening its reach into experiential areas and engaging with customers more fully. The roll-out to date has demonstrated average payback of 16 months, but also that optimum scale is formats above 24 playing positions. Management now plans a roll-out of up to 5,000 BELONG playing stations in the UK. The project is planned to be introduced to Spain during calendar 2018.

Given the number of lease events (272 over the next two years), GMD has significant flexibility to move into larger sites. SD contributes its retail presence and fast-developing estate with increasing numbers of larger stores. Management expects JV sites will have 40-60 positions (we model 40 in openings running at 29 pa).

In our view the key points about the BELONG project are:

The company has identified a number of potential BELONG gaming arenas alongside enhanced GAME retail units. We forecast for five of these to open by July 2018.

The conditions exist for rapid roll-out of the concept, along with GMD’s enhanced stores because the company has over 270 lease negotiations over the next two years, largely allowing for the reorganisation of the UK estate:

This gives management negotiating leverage with landlords because it increases their options. For example, with a small existing store, the company can negotiate with the landlord to extend the property, give a rent-free period etc. Alternatively, the company could form a standalone BELONG arena in a different property, or rehouse the BELONG in a SD development. Or it could convert the existing GAME store into a BELONG arena, taking an entirely new lease for an enhanced GAME store nearby. In a challenged retail property world, this leverage is powerful.

Extension of retail reach under GMD’s control

The creation of concessions in selected SD stores responds to GMD’s and SD’s common customer base. It leverages SD’s UK retail presence of c 500 stores and accords with its property ‘elevation’ strategy of connecting with the consumer through multi-channel experiences and of developing leading flagship stores. But, GMD will have full control of sites selected, terms and unit management.

GMD has now defined its roll-out plans. While these also include BELONG facilities and Game retail concessions within new SD stores, we focus on standalone BELONG units because these provide a clearer model and are not subject to SD’s own estate development considerations. Management expects, over the three years to FY21, to roll out some 4,000 playing stations, although there is potential for 5,000 places. Assuming an average of 40 playing stations per site, this is equivalent to 29 openings a year starting in FY19. Clearly we recognise that there is execution risk associated with a project of such scale and rapidity. However:

For modelling purposes we assume an average BELONG site operates as follows:

The operating metrics used in the model are based on those actually being achieved in the larger size BELONG units, of over 30 stations that form the template for the roll-out. Notably, while the playing revenue is running at £4.50 per hour, total revenue is over three times that. The balance relates to sales of PCs, VR equipment and food & beverage. There is also marketing support from suppliers whose products feature in the arenas, for example tournament and software launch support.

We illustrate below how we model the roll-out to 5,000 paying stations. Conservatively, we assume no price inflation or improvement in utilisation. On these assumptions, the full roll-out takes until 2022 and results in 136 units on the additional assumption that all the playing stations were created in the standalone format. That assumption is unlikely to be literally true, but the economics of the model are unlikely to be disadvantageous if, alternatively, they are housed in a large format SD or GAME retail unit.

Our assumptions result in a steady state EBITDA share of £20m to GMD by FY22.

Interim results showed a decline that was essentially cyclical. EBITDA was down 9% year-on-year, mainly as a result of margin declines resulting from mix changes, mitigated in part by operational efficiencies and cost savings.

The trading period saw a strong retail market both in the UK, up 14.3%, and Spain, up 13.5%, driven primarily from strong sales of Nintendo Switch and associated software, following the initial supply problems in the second half of 2017. Microsoft’s Xbox One X was launched in November 2017, but had more of an impact on the UK market than Spain.

GTV was up by 3.8%, a blend of a 0.3% decline in the UK and a 7.5% constant currency increase in Spain (exchange rates had a c 3% positive effect). The UK had an extreme variation in mix, with console sales up 27.3% but mint software down 1.7%. That software decline reflected in part the 3.9% decline in the number of UK stores, but also a shift to online sales of console product (the offsetting advantage of which is that it releases store space for new activities such as BELONG, PC sales and VR). In Spain, buoyant hardware growth of 16.5% was supported by content sales up 7.4%. On a group basis, hardware was up 24.7% and content up 1.2%.

Within Accessories & Other (A&O), the core category was down 4.7% £78.4m, as a result of console accessories following the hardware growth, while VR sales declined against major launches last year. Events and Esports, still small at £5.4m, were however up 31.5% on acceleration in BELONG and events.

Pre-owned declined 8.6%, reflecting core categories tracking mint software performance in FY17.

Gross profit slipped 150bp overall as a percentage of GTV: representing a 3.1% decline in gross profit.

This was primarily an effect from the much higher mix of relatively low-margin hardware in the period, although individual categories had small variations in their own margins.

Realised savings of £5m in the period mitigated the decline in gross profit. These were spread across property, payroll, procurement and distribution, with a focus on UK retail, where an overall reduction of £4.0m was achieved. However, linked to volume increases, retail costs in Spain increased leaving the net underlying reduction in costs at £1.9m to £101.9m.

The company was very cash generative with operating cash flow of £36.3m compared with EBITDA of £21.2m, helped by working capital improvements of £16.9m of which we expect c £5m to remain in the full year. The most significant cash item was the receipt of £17.1m from the sale of Multiplay in November 2017, following which net cash was up £13.2m year-on-year at £82.2m.

Our forecast reflects our projections as set out above for the BELONG project, combined with expected changes in the market cycle at the longer end:

The market in H218 remains strong both in the UK and Spain, although the Nintendo Switch launch has just annualised, and its growth path in FY17-18 will provide a like-for-like challenge ahead. In addition, although well-established houses have announced Nintendo titles, their timing has not been confirmed, while other game titles have moved from H218 to H119. This has prompted a response from the company, which now has actions in hand to save £6m of costs in H2 in addition to the £5m saved in H1.

Our FY18 EBITDA forecast has reduced mainly as a result of the repositioning of the BELONG roll-out programme. As a result of management’s conclusion that only larger formats should be used, combined with the SD negotiations, we now only expect 24 to be open by the end of the financial year compared with 45 in our previous forecast. But from FY19 we expect the momentum of the BELONG opening programme to drive EBITDA ahead based on our roll-out projections.

The overall decrease in our FY20 forecast needs to be seen in the context of an 8% reduction in market forecasts for that year, as a result of revised expectations for the PS5 launch, while still markedly higher than the previous two years.

Although we have not forecast GMD beyond FY20, our further forecasts for EBITDA of BELONG, against the company forecast to that point should be of interest to investors:

This illustrates the consistent growth of the BELONG concept, which should be largely independent of the traditional gaming cycle, which is governed by the timing of hardware model releases.

We approach valuation on three metrics: peer comparison, DCF and sum-of-the-parts.

There is no exact peer, the nearest being US operator Gamestop, which shares a stated change strategy involving a move away from reliance on physical products. It trades on year one and two P/E multiples of 4.7x, and EV/EBITDA of 2.8x which imply a GMD value of 56p. Also appropriate (as they serve special interest groups) are UK small-caps Games Workshop, Goals Soccer Centres, Everyman Media Group and Focusrite. These trade on an average EV/EBITDA multiple of 11.9x in year one and 10.8x in year two. Averaged between the years, that implies a GMD value of 120p. Taking the average of these two values produces a valuation of 88p (unchanged).

DCF is an appropriate metric because it does take account of medium-term strategy, albeit this is subject to risk across a number of years. In view of that, we apply a high 15% WACC to the future streams. To represent the cycle, we assume the next three years’ revenue growth beyond our 2020 forecast is the same as the average for the previous three before fading to a terminal 2% over the final three years. We assume no change in the 2020 EBITDA margin, and we conservatively reflect maintenance capex at 2% of revenue. On this basis, we value the shares at 71p (previously 72p). That is c 10p sensitive to a 1% change in WACC and c 20p sensitive to a 1% change in the margin assumption.

Since GMD is on a path to realise a business independent of the zero-sum-game cycle, our preferred measure is to create a valuation of each element. We define the growth element as A&O, Gametronics, Esports, Events and Digital. Below the forecast gross profit level, we model specific operating costs in esports, events and digital, and BELONG operating costs. We then allocate operating costs pro-rata to revenue, which is slightly unfair to the growth businesses, which therefore pick up a higher share.

For non-growth operations, we reproduce the cycle by averaging the last three years’ EBITDA of core operations and, after capex and tax, value it as a perpetuity assuming 2% growth, at 4p per share. In the case of growth businesses, we conservatively fade revenue growth from a 2020 rate of 21% to terminal growth of 2%, producing a DCF valuation of 58p. The sum of the two is 62p.

Averaging all three metrics, we define a blended valuation of 74p (previously 93p).

As shown in Exhibit 9 above, the growing operations under the collaboration agreement, as represented by our BELONG projections, is set to dominate the company’s earnings. As shown in our projection at Exhibit 3, the roll-out of 5,000 paying stations would suggest steady state EBITDA of £20m by FY22. On the current All-Share General Retailers 12-month forward EV/EBITDA multiple of 6.5x, that would equate to £130m, which discounted at 10% would suggest an EV of £76m. Subtracting the £45m capital cost of the project gives a value of £31m or 18p. Adding the company’s net cash of £60m (year-end £43m adjusted for disposal proceeds of £17m), or 35p, results in a total of 53p per share.