SPK- Spark NZ (TELCO)

[Snoopy]

Snoopy - were you at the recent ASX presentation asking about Chorus?

'Yes I was' (blatant lie). But since my memory is so poor, what was it I asked about Chorus?

SNOOPY

[Ferg]

No worries......I didn't catch the detail given my mind was elsewhere at the time. IIRC it was during the discussion on Ventia and a learned fellow mentioned CNU in passing .... meanwhile I'm putting 2 and 2 together and getting 5.

Cheers.

[Snoopy]

ChatGPT also doesn't know what it might stand for!

Right, so although I am still stupid on this subject, my stupidity has not yet been usurped by a brainy bot, Good to know. However, you (all) are not off the hook yet. I am setting an alternative puzzle. First a little background.....

I usually read before I go to bed. When I start reading something and I have no idea what is being said, that is usually a good hint to turn the bedside light out. Last night I was reading this:

-----------------------

Fixed - Active Wholesale Income

"Data and connectivity revenue decreased reflecting an ongoing S10 reduction in largely low end enterprise grade legacy products, price competition in wideband fibre products and migration of legacy services."

-----------------------

I had no idea what they were talking about last night and I have no idea now. Can anyone out there read 'telespeak'?

SNOOPY

[clip]

Customers are migrating off older, higher priced, slower connections, in favour of newer connections - both offered by spark and competitors, either way resulting in reduced revenue. I don't know what the 'S10' means however

[Snoopy]

Customers are migrating off older, higher priced, slower connections, in favour of newer connections - both offered by spark and competitors, either way resulting in reduced revenue. I don't know what the 'S10' means however

Thanks for that Clip.

The Samsung Galaxy S10 is a four year old cellphone that "has reached the End of Life and will no longer receive Android security updates."

That is what comes up when I stick S10 into Google. I am not familiar with this phone, or how it fits into the Samsung phone hierarchy. But this doesn't fit with the context of what you have translated Clip. If we are talking in the context of wholesale fixed lines, I don't see how a retail handset can have anything to do with it.

The only connection I can come up with, and it is a tenuous one, is that those who upgrade from a Galaxy S10, (and there will have to a lot of them at once for this comment to make sense) , will suddenly require a higher bandwidth to operate their now more data hungry phones that plug into more data hungry applications. I don't know what these applications might be (a step up in streaming video?). But if a whole lot more users demand live video at once, maybe this increases demand for the highest speed fibre? Thus this supercharges the demand for faster connections?

Of course I may be way off beam with this. It would make better sense if 'S10' referred to some industry standard router that was gradually being superseeded as a 'pinch point' in the fibre network. But what do I know (and on this topic 'very little')?

SNOOPY

[Snoopy]

Snoopy - were you at the recent ASX presentation asking about Chorus?

I am not sure if this is the presentation you were talking about.

https://edge.media-server.com/mmc/p/mnjnxvwk

This is the live version of the FY2024 to FY2026 strategic update. I have decided to belatedly 'attend'. Here are the notes I wrote down as I listened, and the time point at which the information spouted forth:

8:00: '$100m cost saving per year by using wireless broadband'
13:20: To de-commoditise the Spark offering, it is important to hold onto customer relationships, and leverage deep customer insights. Secondly it is important to maintain network capability to become the ideal business partner, aggregating Sparks own products but also a range of local and global providers.
19:33: Multi brand market approach allows Spark to participate in both the value (Skinny) and value added (Spark) end of the market.
23:00; Broadband has limited ability to differentiate and without 'ownership economics' ongoing cost increases put continuing pressure on margins. But Spark can cross sell and up sell within Spark's product portfolio. Cutting fibre input costs, by converting low data use households to 'fixed wireless broadband' can be a win/win for both customers and Spark. Can use data use records to target these lower use customers.
28:20: Small, Medium Enterprise business strategy is to leverage a leadership position in IT. $90m increase in the addressable IT services market from FY2021 to FY2026. Up to 100 small business employees look, Spark will look to deploy standardised packages. More than 100 SME employees? A 'curated IT and hybrid cloud solution' comes into play.
31:00: The introduction of hyperscalism to the NZ market (i.e. Amazon and Microsoft data centres), means Spark has had to change focus from a value chain perspective. Removing back office duplication has been achieved by sacking workers (euphemistically relayed as 'right sizing our cost base'). Spark want to move to 'software defined networks' and move into access and aggregation networks. Hybrid cloud is 'public cloud' (as distinct from 'private cloud'), but with a 'private service layer' over the top. Spark have 35 exchanges and multiple data-centres that can be leveraged to provide geographical resilience and capacity.
36:30: Most growth in last three years for Spark Health has been in 'core services': 'connectivity, ICT, cloud and transformation'. There is a significant opportunity in upgrading legacy systems, and Spark have built some platforms: e.g. 'Kete Waiora', linking health service providers and their patients and providing a patient more control over managing their records. Putting Internet of Things (IoT) devices into hospital beds helps in utilisation and understanding where those beds are. IoT can also help pinpointing the location of defibrillators and how to best access those in real time.
58:30: Exiting PSTN, 3G technologies (out by end of 2025) will provide cost savings. 6G not slated to come on line until late 2028 to 2030.
1:21:30: 80% of Spark's emissions come from electricity.
1:24:00: Revenue Growth Opportunities: Data use low compared to international peers. 5G means faster speeds and more data usage. 5G introduces 'network slicing', private networks as new revenue opportunities. With SMEs, the plan is to move more into the IT services space where Spark have a strong foundation via the hub model. The NZ healthcare sector is one of the least digitised and is going through significant reform, including electronic patient records. Digital identity, through MATTR, is an all new revenue area. These new revenue initiatives will more than compensate for discontinued legacy products, by at least $100-$150m.
1:29:00: Deploying more fibre will allow Spark to use more of their own network and so lower input costs. 5G will allow wireless broadband to grow from 30 to 35% of the market and reduce external input costs.
1:30:15: Investing in the cloud is a way of securing long term customer relationships and long term returns.
1:31:45 : Underlying CAPEX (excluding the proceeds from the TowerCo sale) is a net 10-11% CAPEX to sales ratio (covers maintenance and growth) .
!:33:40 : Net debt to EBITDAI ratio < 1.7 to maintain credit rating. (Note that EBITDAI is short for 'earnings before finance income and expense, income tax, depreciation, amortisation and net investment income.'
1:37:00: Q & A (see Part 2, post 2028)

SNOOPY

[clip]

It won't be referring to the galaxy S10. Something interested I noted in my searches, when looking at the detailed Financials for fy22 and fy23, in both years in the section on cloud KPI's - # of private cloud customers is falling. Isn't that at odds with their investor presentation saying they see growth or are investing further in private cloud? 🤔 seems odd to me

[Snoopy]

The following is my summary of the Q&A session, complete with the timing when each question was asked..

https://edge.media-server.com/mmc/p/mnjnxvwk

I have grouped the answers by topic, rather than chronologically.


Data-centres

1:39:00
Q/ Timing of data-centre(s) expansion program, verses revenue return?
A/ Expansion at Takanini already fully contracted. Will be looking to sign data deals as new builds roll out. Not a 'build it and they will come' program. First expansion at Takanini will come on line in mid 2023. Second stage not until 2026. Spark will give cashflow projections as the build out proceeds.

2:02:00
Q/ How does Spark plan to remain competitive in the data-centre market in the future?
A/ Spark has a first mover advantage and provides much of the connectivity that the new hyper-players (Amazon & Microsoft) will need. There is also an opportunity to partner with these new players.

2:03:15
Q/ 9-10% data-centre return? How is that defined? How do you see it evolving?
A/ Look at incremental returns new data centre investment generates.on a post tax basis. Look at the invested capital that goes into support that. Then divide one by the other. Return comes on stream as the build scales (Takanini due for completion June 2023 is 85% contracted, 100% committed). Data deals are ten year plus time frame contracts with certainty of returns priced in.

2:07:30
Q/ Interrelationship between public and private cloud?
A/ Customers in the Private cloud, are more likely to own their own infrastructure (but not always). A customer can have higher returns with Private cloud, but have to pay for their own capital investment. Private cloud tends to be onshore. A Public cloud, is where a customer does not own their own infrastructure and it is lower margin. The Hybrid approach allows customers to move workloads between public and private depending on what customer needs are: e.g. security, licensing needs.. Private cloud should be thought of more as a product. Datacentres are more akin to a property investment (with a number of tenants).

2:14:00
Q/ How does the Spark view of data-centres being a 'mature market' tie in with the massive investment by the likes of Amazon AWS announcing a $7.5billion investment in data-centres in NZ?
A/ The $7.5b includes downstream external costs so does not all relate to data-centres. Part of that capacity in NZ is likely to be regional (supporting Australia). Furthermore the 'green international marketing tool' of running a data-centre from purely renewable power resources will not have gone unnoticed by AWS. Lot's of data-centre growth in Australia (500MW in NSW alone announced last year).


Strategic Business Direction

1:42.00
Q/ What business areas do you see contracting? e.g. Non-infrastructure cloud?
A/ Broadband market will remain under pressure. Private cloud market had revenue down in the half. Private cloud shifts from a growth business to a more mature business. Therefore we start to manage it for margins, shifting CAPEX to data-centres and at the same time investing in the 'hybrid cloud'.

1:43:30
Q/ Any plans for more co-investment such as with the cell tower sales?
A/ No other asset classes are significant enough in their own right for Spark to want to package as 'for sale' right now. Spark is not like Telstra which has more infrastructure assets available for sale. For high growth markets where shareholders can get a good return (e.g. data centres), Spark intends to keep those 'in house'.

2:18:45
Q/ How to compete with OneNZ and 2 Degrees with their newly announced satellite coverage options?
A/ Satellite for mobile is still early development. By the end of 2024 users 'might be able to text'. By the end of 2025 'some voice capability may become available'. Spark intend to partner with satellite technology providers at a later date. Spark is currently doing more work around enterprise broadband using satellite technology, and a trial with larger enterprise customers who have multi-sites will launch shortly.


MATTR

1:47:00
Q/ What proportion of the $28b market identified for digital recognition does MATTR think it can capture? What benefits are there working under the Spark umbrella?
A/ No dollar answer on what part of this international market can be snared, but it is a multinational growth opportunity.. Deep tech and R&D cycles are quite different to an operating business. A strong backer gives MATTR freedom to do the work they set out to do.

1:53:00
Q/ What CAPEX is it projected will be needed to develop MATTR going forwards?
A/ MATTR started with just 6 people four years ago in an evolving market. Resources MATTR has now is entirely appropriate to the level of market maturity and to keep MATTR in a leadership position. A key differentiator for MATTR is interoperability.

2:17:00
Q/ Does MATTR use blockchain technology for verification purposes?
A/ The architecture can accommodate the use of distributed ledgers but it doesn't need to. Most clients do not require these for the solutions they are embarking on. MATTR can achieve the properties of decentralisation, anchoring into existing groups of tract and decentralised Public Key I infrastructure (PKI), without needing to use blockchain at all.


Financials

1:44:30
Q/ What is the outlook for growth above inflation?
A/ 5G! MATTR is another growth opportunity. Thirdly, selling capacity already built into data centres (most costs are incurred before the income stream builds)

1:49:30
Q/ Growth cost reductions, projected as about $60m per year. = Business as usual?
A/ Any planned 'cost out' is allocated to very specific areas, together with a detailed program that drives costs down on an ongoing basis.

2:20:15
Q/ What is the free cashflow growth split between the 'new investment space' and 'core development of existing business'?
A/ Majority of three year projected cash-flow growth is coming from core operations. The $350m of incremental investment announced doesn't really come on line until FY2026, and even in the data-centre space they have to scale some income over time.

2:20:15
Q/ Dividend imputation guidance over the next three years?
A/ Dividend guidance is given annually. But Spark's aspiration is to grow the dividend over time and the ambition is to have that fully imputed.


5G mobile

1:51:00
Q/ 5G spectrum availability and capacity?
A/ Nearing end of spectrum allocation discussions with the government. Expect an announcement in months/weeks. mm wave still pending. 600Mhz helpful to overcome rural topography. NZ more spectrum rich than most other countries. Coming out of 3G, Spark can reallocate that spectrum. 5G network can be continually densified over time.

1:56:30
Q/ What is the addressable market for Wireless Broadband (Spark mentioned 35% goal by 2026)?
A/ 5G will drive this.Also understanding of data capability of households. The 35% is an interim target and use patterns out into the future will determine any targets beyond that. The relentless upward price of Chorus broadband gives the opportunity to adjust wireless prices upwards for some customers.

1:58:30
Q/ 5G monetisation strategy, more detail?
A/ Monetising gaming experience - low latency with 5G standalone ideal for this. Also entertainment more broadly. There are annual price reviews, and further opportunities for wireless broadband and the enterprise market.with convergence of technologies.

2:15:45
Q/ Functionality of 5G tied to the public cloud, not on premises equipment?
A/ Different vendors play in the 'stand alone' 5G space. Hyper-scale databases with a standard vendor is one option. But one can use a standard vendor with the computing power of the workspace.


Customer management

2:09:00
Q/ Cybersecurity?
A/ Spark has a large team, numbering 180. This includes a 24/7 security team fighting DDOS global attacks, but also people who work in 'security response' and governance, and 'security infrastructure'. These are people across all different parts of the business. Security is across all phases of designing and building and from operational perspectives. What happened? How did it happen? And what can we learn from this? are the questions asked when security is breached. This is a culturally evolving space with input from overseas experience (security issues often are seen overseas before they pop up as a problem here.) Spark even work with the GCSB.

2:09:00
Q/ How long is customer data stored?
A/ Spark's privacy policy very specific about what data Spark collect and use. Security is built into all systems, Customers will continue to be happy to share it with Spark if they can see the benefit. There are too many data types to go into individually. Spark use AI modelling in which they input customer data to enhance the customer experience. But Spark do not store data they do not need that might create bias e.g. gender data. Furthermore, Spark has a 'digital trust team' that provides overarching policy on data collection.


Spark Personnel

2:18:00
Q/ Skills training partnerships to combat lack of IT skills in the workforce?
A/ Spark are already working with a range educational institutions like Te Pukenga (the old polytechnic colleges combined into one) , plus mirco-credential certifications with partners in the technology and product space. Rethinking variants of these partnerships concentrating on the internal skills Spark needs to bring business opportunities to life.


SNOOPY

[Snoopy]

The following is my summary of the Q&A session, complete with the timing when each question was asked..

https://edge.media-server.com/mmc/p/mnjnxvwk

I don't know whether anyone else on this forum tuned into this presentation. It was nearly two and one half hours long after all. I had never seen CEO Jolie Hodson in action before, but in that Q&A session at the one hour forty mark, I was mightily impressed. In fact I had never seen anything quite like it.

Jolie was on the stage with her senior management team, ten people I total I think. Then various analysts were firing questions at her, and she would say something like.

"Well that is actually three questions which I will split up like this. Senior Manager X (SMX) will answer this bit. SMY will answer this bit. and SMZ will answer this bit."

To have the instantaneous brain power to deconstruct a 'surprise question', and do that kind of delegation in an instant, I think shows someone who is extremely smart and right on top of their game. Jolie obviously had total confidence in all her senior management, although if they were to stumble Jolie has enough overall business knowledge to 'keep them on track'. When one question started to get a bit technical Jolie interrupted her SM and said, :We will finish the answer privately later", and the Q&A rolled on. It was a very tightly run Q&A, yet everyone got their answers.

If you want to know how to run a Q&A session, this is a very good example of what you should do. But even then, you would be hard pressed to get someone of the caliber of Jolie Hodson to chair it. I think we shareholders are pretty lucky to have Jolie Hodson as Spark CEO.

SNOOPY

[Snoopy]

The two 12.5c dividends paid during FY2020 were only 75% imputed whereas the two 12.5c dividends paid during FY2021 were 100% fully imputed. A 75% imputation rate means that 9.375c of the dividend is fully imputed and 3.125c is not imputed. So to make that 75% percent imputed dividend 100% imputed, we need an extra 3.125cps of fully imputed earnings to be booked every six months (12.5c dividends were paid twice per year). In fact the incremental earnings between FY2020 and FY2021 went down from 0.9c to 0.6c. So it looks like changes to earnings from regulatory review, substituting for Chorus services and Covid do not explain the extent of increased profitability of the company.

AP2021 announced the FY2021 dividend payments of 25cps (AP2021 slide 3) were fully imputed, supported by resilient free cashflow. That is an odd thing to say because fully imputed dividends must come from actual fully tax paid profits. 'Cashflow' does not guarantee that! Unless, that is, Spark are still using free cashflow to pump up the tax paid in advance to give the impression that the increase in taxable earnings is sustainable!

I am looking at this from another angle, starting with 'the facts we know'.

"Spark had an imputation credit balance of nil as at 30-06-2020" (AR2020 p94, confirmed in AR2021 p100)

A Nil imputation credit balance means that all of the imputation credits paid out after 30-06-2020, must have also been paid up after 30-06-2020. So in the two year period following EOFY2020, what fully imputed dividends were paid out?

Dividends Paid FY2021 & FY2022	Gross Dividend	Net Dividend	Imputation Credits
Second HY dividend FY2020	$319.4m	$230m	$89.4m
First HY dividend FY2021	$320.8m	$231m	$89.8m
Second HY dividend FY2021	$323.6m	$233m	$90.6m
First HY dividend FY2022	$325.0m	$234m	$91.0m
Total			$360.8m

Calculation Notes

1/ Gross dividend = (Dividend Paid)/0.72, Imputation Credits = Gross Dividend - Net Dividend



-------------------------

Now, how does that $360.8m imputation credit total paid out line up with the amount of tax paid up by the company over that same period? From the respective cashflow statements:

Income Tax paid over FY2021	$188m
Income Tax paid over FY2022	$160m
Imputation Credit Balance Owing EOFY2022	$16m	(See AR2022 p120)
Total	$364m

This means enough tax has been paid to cover those fully imputed dividends. Or perhaps more correctly, enough tax would have been paid, if Spark had got around to paying that outstanding tax debt of $16m of imputation credits owing! But as nztx pointed out in post 2004, the imputation credit balance isn't really owing (yet), because of the mismatch of the ending of the Spark reporting financial year, and the Spark tax financial year that ends 9 months later. That means the unpaid imputation credit balance of $16m is actually an accounting construct, caused by Spark choosing to put a line under their year at 30th June, while the same $16m is a bill 'not yet due' from an IRD perspective.

SNOOPY