Reality check: Faster NBN shaping won’t bankrupt ISPs


opinion Over the weekend, I received a flood of outraged comments accusing me of not taking the costs of ISPs into account in my article exploring alternatives to the current speed shaping models being proposed by ISPs for the National Broadband Network.

At the heart of the matter is my belief that it is ridiculous for ISPs to simply apply the same shaping model used on ADSL networks to their NBN pricing plans, when the top-line speeds available under the NBN are at least four times than available under ADSL2+. I believe that it is unfair to shape NBN users down to 512kbps, 256kbps or even 128kbps once they have exhausted their quota.

In this context, I called for shaping models to be re-examined and proposed a number of alternatives. However, a number of readers immediately cried foul.

“If you want to leech, PAY FOR IT!” screamed one commenter. “Plain and simple!!” Another, Nic Crouch, said the point of shaping was “to frustrate you” into signing up for a higher value plan. “Don’t like it? Get a bigger plan,” Crouch wrote. “This is a very egalitarian socialist approach which simply doesn’t work in the real world,” was the way another reader characterised one of my proposals. “Renai, you and other journalists don’t seem to get it. The NBN is not equal to the Internet, and isn’t going to make the Internet cheaper — in fact it is going to make it more expensive for ISPs to deliver,” wrote a fourth reader.

On Twitter, former Internode network engineer Mark Newton went further, claiming the presence or absence of the NBN made “no difference to the economic reasons for shaping” and that for users, the best option was to buy a quota that was big enough that they never got shaped. “How is it in a user’s best interests to have their ISP perform badly and go broke?” Newton asked.

The general thesis of most of those who criticised the article was to argue, essentially, that customers are relatively stupid but still capable of making their own decisions. To massage customers into paying the right amount each month for the amount of data they use, so the argument goes, ISPs need to apply both a large stick (shaping customers’ speeds) as well as a carrot (the lure of increased quota on a higher plan). The side effect, of course, is that most customers will be paying for at least some quota that they don’t use, as they come in under their monthly tally on a regular basis.

Now all this is very well and good. I don’t disagree with this argument in general, and I also agree that care needs to be taken by ISPs to ensure their businesses are sustainable in an NBN world. However, as often happens in Australia’s telecommunications sector, I am also concerned that those making this argument may have locked their minds into a little box and are refusing to consider that alternatives to this carrot and stick approach already exist.

To illustrate this argument, let’s compare the ADSL broadband plans offered by several of Australia’s largest ISPs, TPG and Internode.

Internode’s approach to broadband pricing is clearly the ‘carrot and stick’ model described above. The company gives customers complete transparency into its costs, offering them different pricing options depending on whether they use its own ADSL infrastructure or that of Telstra or Optus, and it then offers a large number of tiered quota points to choose from. If a customer exceeds their quota, they can easily go up to a new plan, or buy a ‘data block’ to tide them over until the end of the month.

This model clearly assumes a certain degree of intelligence on the part of the customer, mediated by the ‘carrot and stick’ shaping/upgrade model to keep the customer within the bounds of Internode’s profitability. A worthy compromise, you might say.

Now let’s turn to TPG.

Unlike Internode, TPG clearly assumes that its customers are pretty stupid. The company does not provide much transparency into the costs which make up its broadband services, and it appears to prefer that customers simply let it make its choices for them. The company consciously guides customers towards its ‘Unlimited’ plans with no quota restrictions by pricing them the same or very similarly to its plans with discrete quota amounts. Dozens of plans? Data blocks? Whatever. TPG customers can just set and forget when it comes to broadband.

So what’s going on here behind the scenes? It’s very simple.

I haven’t talked to the company about it, but I am betting that TPG has used its many years’ worth of experience in Australia’s broadband industry to build up very complex prediction models on how much quota its customers soak up monthly. It has then applied that figure to its customer base as a whole and calculated how much it needs to charge each account monthly in order to continue to make a profit. That figure, combined with a home phone, appears to be something close to $60 for TPG — whether you bundle or don’t bundle.

Now, on the face of it, this appears ludicrous. There is no doubt, after all, that a company like Internode would baulk massively at the idea that it could make any kind of profit at all from charging customers $60 a month for unlimited amounts of broadband. On an individual customer profitability basis, this just doesn’t make sense.

But it’s important to realise that what is going on here is cross-subsidisation. TPG has likely calculated that a huge proportion of its customers do not download more than something like 20GB a month, and that it will make more than enough money from them to subsidise that small proportion of customers who download something closer to a terabyte. It’s a nice little trade-off which allows TPG (and others like Dodo) to bundle up this calculation into the sort of ‘Unlimited’ marketing offer which will give customers complete peace of mind. Quotas become an anachronism of the past and customers just use what quota they use, without worrying that they’ll run out.

Meanwhile, TPG makes profit. A lot of profit. In the last financial year (PDF), the company chalked up net profits of $78 million off revenues of $574 million. Not bad for a year’s work.

What this little example illustrates is several points which are very important to the NBN shaping debate. Firstly, the ‘carrot and stick’ model which focuses on the individual profitability of broadband customers (and stimulates the need for shaping) is not the only model possible. Cross-subsidisation is another very prevalent pricing model in Austraila’s broadband market, and there is no reason to believe it will not exist in one form or another in an NBN world.

This goes some way to implying that higher shaping speeds ought also to be possible in an NBN world without sending ISPs bankrupt, as decently improved shaping speeds (say, 5Mbps as an example) will be in the realm of existing lower ADSL speeds.

The elephant in the room, of course, is the fact that ISPs do not yet have enough customer data to make the NBN cross-subsidisation calculation that TPG and Dodo already have with respect to ADSL. The NBN will unlock vastly higher speeds and better latency, and ISPs will not be able to plot average customer behaviour on a graph — and profitably allocate averaged prices — in the area for some time to come. It may be that the average customer will download dramatically more using their NBN connection than on ADSL — and that ISPs’ ability to soak up data costs will be diminished as a result.

What this lack of data will likely mean is that in the medium-term — say, the next few years — most NBN plans will focus on the ‘carrot and stick’ model, until they get better information. However, I am betting that after that point, as the NBN pricing market starts to mature, cross-subsidisation models will emerge, and strongly so. The gradually decreasing cost of international and domestic backhaul traffic will also aid in this trend. There are a lot more fibre cables being built around Australia and from Australia right now, with the effect that capacity will gradually become cheaper as more competition enters the market.

When this happens, the existence of cross-subsidised plans will likely force up shaping speeds for those using the carrot and stick model as well. When you’ll eventually be able to get an ‘unlimited’ NBN plan, those ISPs peddling tiered quota packages will find it harder to justify radical shaping speeds of up to 512kbps. I think it likely that to retain customers, they will need to drive shaping speeds much higher.

Thoughts? Queries? Extended rants? The comments box is your oyster :) I would particularly welcome comments from ISPs on my reasoning here. As I’ve previously mentioned, as a journalist I deal primarily with top-line trends. I do not have the knowledge to make this argument down to the exact dollars and cents level ;)


      • TPG hasn’t changed anything about the pricing about it. That was the main concern when they bought PIPE’s.

      • Point is, that internally for TPG, their bandwidth cost is close to nothing, where as externally they charge extra (this is obvious)

  1. It seems you would happy if there were more usage options available? ie. if the plans currently offer 200GB and 1TB usage, and you want to use 500GB, you want a 500GB that is priced between those two.

    Basically you’d like the ability to select your own speed and usage and have a package taylored to your needs, correct?

    • What I want it some more innovative thinking from ISPs on the issue of NBN pricing … not just rehashes of the same model, which has its problems :)

  2. You are assuming TPG is passively collecting data on how much their customers use, and cross subsidizing to match. The reality is they are much more active than that. Have a look at complaints about TPG peak time performance – TPG *controls* the amount of data its users consume by under provisioning its links for peak times. They set a price and a profit margin, then assign bandwidth accordingly. This is no different from shaping, except that it operates without user knowledge. Why do you think the ACCC is so hot on ISPs being forced to advertise the actual peak speeds on NBN, not theoretical max?

      • It’s misleading is what it is. I get that there are things outside of an ISP’s control that will affect my download speeds, but it is not acceptable for an ISP to throttle my downloads when I am within my quota.
        I’m surprised that ACCC haven’t taken TPG to court for this sort of crap. At least with other ISPS ‘unlimited’ plans, they tell you in black and white that you get XXGB at a maximum speed of XXmbps for XX dollars and then you get shaped once you have maxed out your quota.

        • “they tell you in black and white that you get XXGB at a maximum speed of XXmbps for XX dollars and then you get shaped once you have maxed out your quota. ”

          Bs they tell you the max speed you never get then you get slowed down at peak time anyway (Probly slower then tpg shape you) because of who really knows…..

      • Pretty dishonest approach really. Read their threads some time. The things they try to blame evening slowdown on are rediculous. They must be the only ISP that has a lot of customers thinking their evening slowdown is probably because of their neighbours electric hot water system. I’d never go to TPG if I wanted evening speeds, only if I wanted to leech.

      • I suspect that the ACCC would consider it a deceptive and misleading business practice rather than smart.

        TPG get away with it due to the customer demographics, most of their customers don’t know any better.

        I suspect there are a couple of key reasons TPG are able to offer the plans they do…
        1. They under-provision their network
        2. They own PIPE
        3. Telstra cut them a sweetheart deal to skew the market

        Also, you really think TPG have better customer data than iiNet or Internode?!

        If I offered iiNet or Internode complete national data for all broadband users in Australia it would be of limited use to them. They need to have the statistical modelling for *their* customer base, not someone elses.

        No one understands their own customer base better than the actual company itself.

        • Also, you really think TPG have better customer data than iiNet or Internode?!

          Never used Internode, but from personal experience iiNet flogs TPG when it comes to phone support.

          During my time with TPG I called them a couple of times, extremely painful and one of the reasons I left (no point in staying with an ISP who draws what should easy to resolve problems out).

          So far I’ve called iiNet once, everything was a breeze, and the fault (the TPG issue was billing, this was actually a fault), was resolved within a few days, and they kept messaging during that time with status updates.

          TPG may be cheap but I’d never go back to them, and it would take a bit for me to leave iiNet right now.

    • TPG and Dodo both run networks with peak-time congestion. They simply don’t have enough bandwidth to go around if a large number of users start using up their unlimited allowance. These ISPs offer enough rope to their users for them all to strangle each other, and so that’s what often happens.

      With ISPs that use the “carrot and stick” approach, they will always have some spare capacity. They have done the sums, enforced limits on users, and created a network that offers the same fast speeds, night and day, with no congestion.

      It’s up to users to decide what they want. Do they want a cheap two-lane highway that works most of the time, but experiences traffic jams if too many people hop into their car at once. Or do they want a multi-lane freeway that is never congested, but has an extra toll on people who drive back-and-forth all day.

      Personally, any downtime or frustrating congestion is enough to turn me right off an ISP. I’m happy to manage my usage or pay a bit more my quota, and to get a connection that just works. All the time.

    • As far as I know, the issues that TPG had with slow downloads at peak happened when TPG first released unlimited and hence they needed to upgrade their infrustructure to handle the massive surge in load (upgrading routers, fiber links etc etc)

      As far as I know of, and currently (not when unlimited was released), TPG does not do any shaping.

      DODO is an ISP that is known to shape P2P/Torrents, and that is stated in their contract

  3. Based on what you are saying is that maybe the reason that TPG are not offering services on the NBN? Because they can’t actually offer unlimited?

  4. There are so many elements to this. Firstly the declining average cost of transmission doesn’t translate to a declining marginal cost of the transmission used by the heaviest user. But secondly the important question you raise is what to do when the “limit” is reached. Shaping down to an unusable level (and that’s what your average broadband user thinks of current shaping speeds) is really about as effective as disconnection. There are in the market different options as you note – force the customer to have to elect to change plan or the more palatable “add on data” (first intro’d I think by vividwireless….).

    To that you are adding the suggestion that as speeds increase the shaped speed should increase to be as effectively useable as shaped speeds used to be. But it does open up additional possibilities, such as “add on data that increases the shaped speed but doesn’t take you all the way back.”

    The challenge is in keeping the offering simple enough for both the customer to understand and the IT and network systems to implement. The latter actually tends to be more the limitation on the creativity of marketing than the assumption of the intelligence of consumers.

    Finally I’m not so sure the regulators would agree with you about the assumptions of the intelligence of consumers – they think consumers already don’t understand current market offers.

  5. TPG makes a lot of profit for offering substandard services and useless tech support.
    Shame when I was with TPG they were excellent now they are just crud.
    They don’t even offer Off net Unlimited ADSL plans anymore not even on ADSL1 speeds.
    If TPG can’t handle their own ADSL god help them with the NBN.

    • Useless tech support?
      Hardly, they have 4 active members on whirlpool which swiftly respond to issues.
      I have had a few issues with TPG and they were all resulted by Karen C on whirlpool (in the end it was telstras dodgly line for most of them).

      I believe they are still a good ISP, there are thread that popup that thank TPG and it gets many thanks. The customer that complain are the loudest in the case of TPG.

      • They swiftly respond to issues that are obviously a bad line or an easy fix. They swiftly give anyone who is having congestion problems the run around. “Post stats here”, “I have 6 times”, “”So it only happens in the evening? Checked your flourecent lights, do the neighbours have an electric hot water heater”, anything but admit congestion on TPG.
        Your lucky you had a bad line, if it was that your data had to crawl through on of their underprovisioned links you’d still be trying to get them to fix it.

  6. The problem with TPG unlimited vs high shaping is that you get large numbers of customers who download very little on TPG unlimited.

    If someone gets shaped, then they are *not* someone who downloads very little. So, the usage pattern of someone shaped from 100Mbit to 10Mbit because they went over their 200GB quota is very, very different – and the ISP must *always* have 100Mbit of CVC spare to ensure that their 100Mbit customers do not encounter congestion.

    Compare this to TPG unlimited, which is *only* offered in certain areas where backhaul is very cheap (also, they recently stopped offering it over 8Mbit Telstra ADSL), and they can “get away with” smaller amounts of spare capacity due to lower ADSL2+ line speeds.

    Then there are the “tricks” that TPG allegedly uses in peak times. I want a service that always performs well.

    The whole debate is moot – if you get a big enough plan, you won’t get shaped. If the demand for plans bigger than 1TB comes along, then at least one ISP will offer such plans.

    • If noticed in peak times on my exchange that i get around 650 kbps which is the same as i normaly get. (im on TPG unlimted in Eltham, not a major town)
      In the morning i can hit 740 (which is the fastest my line can go)

      • My TPG link always gets between 1.2 mb/sec and 1.6 mb/second, regardless of the time of day. (honestly; I use it mostly from 5-10 when home from work).

        Cheltenham area, melbourne. (all ports full in exchange last I heard)

        Obviously my experience is not that of others.

    • TPG unlimited is covers like 60% of Australia at least, PIPE is a international fiber network. Its mainly in rural areas, where PIPE backhaul doesn’t reach the exchange (and hence ULL/LLS can’t be used) that TPG is forced to go through Telstra

      TPG is actually the ISP that has the 3rd highest amount of DSLAMS installed, after Telstra and Optus

      They are doing better then ISP’s such as iiNet and Internode in this regard

      • Think you mixed up iiNet with Optus.

        It’s Telstra – iiNet – TPG – Optus – Internode iirc.

        iiNet only recently passed TPG for DSLAM count, they are both kinda neck and neck.

  7. it does depend very much on who you are with and their business view of the world as to the current situation.

    TPG certainly do cheap out on provisioning/contention so they pay the minum cost they possibly can – yes it makes them a ton of money but ive much preferred the Internode approach for a long time, i definitely prefer an ISP willing to consider its customers arent complete morons to be led about by the nose.

    as to the shaping it may become a moot point in the end anyway. with the speeds and capacities on offer from the NBN the only real way for data caps to move from here on out is up. i think there will come a time where this plays to the extent that shaping will become a power user thing, where heavy users keep bumping into their monthly limit.

    thing is, well im looking at the Internode NBN second release pricing which gets you a floor of 30 GB a month and immediately goes to 300 gb next tier.

    with the natural data inflation effect the bottom tier will inevitably come up as more cables get laid and offshore haulage as a group – whichever cable used – comes down. If internodes next revision comes at 100GB for the floor, say, there will be many many users who simply wont get within a bullsroar of that use and for whom shaping is a moot point. it doesnt matter if its 128 KB or 512, use wont reach the trigger so those speeds will not be going to come into effect.

    so what happens when the floor tier is 300 GB? for all but inveterate Leeche -imean, linux distro folks, even if you can reach your cap in a matter of hours it isnt likely to happen when all you do is check email etc. so data download usage, the driver of bill shock which shaping was created to deal with, will go away as a concern for a large part of the market. the people who will still live in a shaping world are the hardcore line users, trying to screw every cent of value out of their connection every month. for those, i think the carrot option will disappear and the stick will become more evident.

    i do think IPTV will bring up the average user level of data use but only as a casual use thing for the less net savvy of us. i would be surprised if Joe Average NonTechGeek does 100 GB of fetchTV the first month after being introduced to it, say. so i think there is room for the floor tier to come up, with room for new NBN uses but for most but the hardcore line users, no real need for shaping as a cost mitigation strategy.

    we arent at that point yet so shaping will still be an anachronism until then – no argument there. but i think the whole idea that this is a ‘problem needing fixing’ is overselling the issue a bit, and personally feel natural market directions for the next 5 years will likely leave shaping as a nonissue for most.

    • why? Just because we don’t always agree with him?
      Opinion pieces are just that. They also encourage debate and at least Renai will actually reply to some of the comments.

      Oh god. Now I sound like a delimiter fanboi. :P

    • I know what you mean, Bob, but the difficulty is that I realise my readers often know a great deal more than I do in this area. It’s impossible for me to be an expert in every field. I do, however, have a skill for provoking debate. During that debate a great deal of information often gets pushed to the top of the discussion — information it would take me a lot longer to source myself.

      I wish I had more resources and more time for more research … but the reality is that I don’t ;) In the meantime, I do what I can :)

      • Most of it would be BS anyway. The number of people who think they know everyones business and all their (confidential) agreements astounds me.

  8. Renai, I think you are being deliberately dense here. You have still not addressed the issue of ISP costs (and lets face it,reducing ISP transit costs is the only reason that ISPs shape at all) aside from asking for ISP’s to think ‘outside the box’. You are entitled to ask the question, but to be so passionate in the face of the obvious is pretty frustrating.
    Whilst I would love to see quality, unlimited and affordable broadband options, it is simply not realistic with the current state of the market. I would also love to be able to buy unlimited amounts of petrol for $10 a month….

    • You hit the nail right on the head . Pay for $10 of petrol, water, electricity or anything and you get to use the full $10 ammount.
      ISP’s and Phone companies are able to gouge their customers two fold. Thru un-used quotas that expire within some time limit; and thru speeds “up to…” that are never realized.
      Caps and Quota plans arer a rip-off. Charge customers according to what they consume.Factor in yuour overheads into that price and we all have equity and fairness.

      • Maybe a model like electricity is what we need. A flat charge for ISP access and all data billed by volume. Maybe a sliding scale for volume rate. On peak/off peak tarrifs.

      • The problem with this option is the amount of data that could be consumed:
        * 12Mbps = ~130GB/day
        * 25Mbps = ~270GB/day
        * 50Mbps = ~540GB/day
        * 100Mbps = ~1TB/day

        Unlike the petrol, water, electricity, gas, etc. data can be consumed without being aware of it. Someone is infected by a virus; learns about bit-torrent; runs skype with sharing on; etc. and suddenly massive bill.

  9. I think Renai is right. As at June 2010, the average ADSL plan cost $50 for 50 GB, and the average data consumed was 7GB. ISPs with a large user base could market a 50GB allowance, but cost it as if it was 7GB, because a few leeches would be cross-subsidised by thousands of others. For now, NBN retail service providers simply do not have the scale of users to risk such pricing, and will need to cover their costs for a while until they do.

    A few years back, when most folks had 2GB or 6GB allowances, Internode launched “Easy Broadband”, offering 50GB for $50. The idea was that this allowance was big enough that few users would ever hit their cap, and that is exactly what happened. Those who wanted to download the whole InterWebs onto their network drive every month went elsewhere, which was a relief to everyone else. The same will happen on the NBN, except that many more providers of all sizes will be able to build the scale needed to cross-subsidise heavy users with more numerous light ones.

    • Unfortunately the 121 POIs model actually makes it harder to build scale on a national basis. Indications suggest most expensive part of the network is the NBNCo CVC charges. This means that bandwidth needs to be managed at the POI level.

  10. Heh, to dispute the first option in the other article, just do the math. 100mbit getting shaped to 10mbit does not solve the problem that shaping exists for – to limit a user that goes over their “budgeted” quota so they can’t cause problems on the network.

    10mbit might sound like a much smaller amount than 100mbit, but the reality is someone limited to 10mbit can still download ~3TB of data. It doesn’t make sense given even the heavy usage plans are only in the hundreds of megabytes.

    Traffic shaping is simple, easy to understand and effective. I’ve never been shaped because I have a plan that is slightly more than what I need.

    Bandwidth isn’t an unlimited resource; more importantly it is a SHARED resource, so people abusing it create a negative effect on others.

  11. Just to add fuel to the fire you really need to compare when people actually use the data not just how much they use.

    If you’re able to look at an ISPs overall usage graph it ramps up during the evening peak times but drops off in other times, including offpeak times.

    So although someone may download say 800GB in a month, if they do 95%+ of that outside of peak hours then their overall impact to the network and the requirements in infrastructure to meet their needs would be less than a 50GB user who does all of their usage during peak hours, quite simply because they are doing their usage when the network is under utilised with ready bandwidth available.

    So another question is whether or not a blank usage count is even enough, should the usage be affected by times of day, like for example the harbour bridge/tunnel toll is costed based on time of day (aka congestion)?

  12. The NBNCo Corporate Plan – has this to say about Connectivity Virtual Circuit Pricing on page 103.

    “The Connectivity Virtual Circuit (CVC) in the product construct is an aggregation point where the Access Seekers can choose to contend their traffic to create differentiation. CVCs can be used as proxies for usage charging. The CVC is purchased based on the bandwidth required to service the total amount of End-Users being aggregated across this link.”

    From this I can only assume that NBNCo are expecting budget RSPs to purchase insufficient CVC and pass those savings onto consumers, while premium RSPs will provision sufficient headroom for peak periods.

  13. The NBNCo Corporate Plan – is interesting reading about expected usage of quotas on page 105, where NBNCo discuss estimated retail pricing:
    “The first bar of NBN plans, represents ‘low use’ End-Users based on a 12/1Mbps 50GB plan with an assumed average download of 18GB. The second bar represents ‘medium use’ End-Users based on a 25/5Mbps 200GB plan with an assumed average download of 40GB. The third bar represents ‘high use’ End-Users on a 50/20Mbps 500GB plan with an assumed average download of 75GB.

    Notice how they expect on average only 75GB to be used for 500GB plan (only 15%!). Compare with the many threads on whirlpool about how to use up the last GB of quota before the end of the month.

    To understand why Internode’s flatrate failed and why I think quotas are an important management tool is the “tragedy of the commons” needs to be appreciated. Quoting from wikipedia -
    The tragedy of the commons is a dilemma arising from the situation in which multiple individuals, acting independently and rationally consulting their own self-interest, will ultimately deplete a shared limited resource, even when it is clear that it is not in anyone’s long-term interest for this to happen. This dilemma was first described in an influential article titled “The Tragedy of the Commons,” written by ecologist Garrett Hardin and first published in the journal Science in 1968.

    • I guarantee every leech has a Whirlpool account, but fewer than 5% of pensioners do. Many pensioners are currently on perfectly adequate $29.95 5GB Telstra ADSL plans with their $22.95 Home Line Basic rental. When their current $52.90 monthly bill (before call costs) is replaced with an NBN plan of about $50, most will continue to consume modest data. My point is that NBNCo’s 15% prediction of quota usage is probably a very well-founded estimate, with a mere few thousand leeches out of ten million connections.

      As for the NBNCo Corporate Plan 2011-2013 forecasts in general, there was a political imperative not to overstate the revenue expectations that will repay the build costs of the NBN, therefore all of the forecasts were conservative. On any extension of historical growth charts the takeup and speeds sought should exceed those forecasts. (My view is that the build will pay for itself well before forecast, and a fire sale of the infrastructure would be folly, as it is a permanent cash-cow for the public purse if kept.)

      Of course data usage will skyrocket when bandwidth grows, and competition will ensure that myriad combinations of caps, throttling, unlimited-but-underprovisioned offerings appear on the totally competition-driven landscape. In the meantime, I applaud Renai for kick-starting this discussion about truth in advertising and sustainable pricing of plans.

      • These pensioners would be part of the 50% that NBNCo are predicting (page 118 of NBNCo Corporate Plan) will sign up to the 12/1Mbps plan.

        Consider this for that $50 spend, the pensioner could ditch the NBN and go with their wireless operator $39 is the cost of an unlimited calls plan and 4GB of data. I wonder if the forecasts are conservative enough, especially with the launch of LTE meaning that LTE plans will be faster for customers in the right locations.

          • there’s that same old, tired 50% – 12/1 tune again…

            i think you mean FTFY for the umpteenth time?

          • We just pick different data points on the chart. 36% connect at 12/1Mbps in 2028 in my opinion is even worse that 50% now.

          • We just pick different data points on the chart.

            Which is very convenient if you are trying to deceive people. The big problem is that you keep citing the same graph on page 118 without even mentioning the point where you picked the data. I’ve looked at the whole graph not just a small chunk.

            So +50% for 12/1mbps is up until 2017 after that 12/1mbps is no longer +50%. After that the majority are on higher speed plans. It’s not hard to understand but you keep omitting that information. The prediction goes up to 2028 not 2017.

            You know what else is funny, you thought 50% was up until 2028, you were picking that same point on the graph too, it’s only after I pointed it out your error that you changed your line. Unfortunately it means your argument is meaningless.


            on page 118 there is a chart showing that 50% of customers will connect at 12/1Mbps for the foreseeable future (2028)

            Seems the foreseeable future does go up to 2028 after all lol.

  14. I like your comparison of Internode vs TPG for what is actually shown on their websites. Internode’s website is incredibly hard to find the information you want, its a complete information overload as well as stupid terminology they’ve made up for the different plans. I’m an electrical/telecoms engineer that looks at datasheets most of the day, and still have to do some research as to what they actually offering!

    • electrical/telecoms engineer

      Do you know your ABC’s
      Hope your employer don’t see your post i would sack you

  15. As long as there is a quota limit, some people will try and use all of it because, in their mind set, if they don’t use all of the quota they pay for, they are losing. Make the quota unlimited and they lose that focus, that maximum figure to aim for. Personally, if I was on an unlimited plan, I wouldn’t use a GB more than I do now. I realise some others want to senselessly download the internet but they would be a minority that could be dealt with as deemed appropriate by ISPs. Even though a plan is unlimited, that’s not to say there couldn’t be a theoretical upper limit imposed on those who constantly abuse their connection.

  16. Renai, Shaping is a ‘free’ alternative to ‘No more internet for you!’. You are proposing that ISP’s give you ‘free broadband’ when you have used all of the broadband you pre-paid for. That’s Nuts and it’s not going to happen. Imagine charging an ADSL customer $50 for 50gb @ 8mb, and giving the same thing away to an NBN customer as a reward for using up their quota.

    Reading between the lines, the basis of your argument against shaping seems to be ‘my 50gb costs $50, why can I not have 2GB for $2???’. Data costs are not that transparent, and no amount of wishing will make it so. For one, todays plans are designed to make you over-purchase to avoid running out of Quota. I think that’s a particularly shitty model. I like that you’re looking for alternatives.

    You could argue to have a retainer + excess charge model that is comparable with high volume plans (say $30 Rental and 40c per GB). This hasn’t been done because:
    – a) there is no reliable way to predict demand, and therefore, to meet bandwidth demands before congestion occurs and
    – b) assuming you solve a, the cost per GB is determined over a period of time which concludes after the usage occurs, which means customers don’t know what rate they are paying for their usage as they use it
    – c) customers currently over-purchase, allowing the ISP’s to advertise attractive plans that will only get used maybe 40%(?). Move away from the cross subsidy model and you need to bump up your per GB rate to actual cost + margin, so these usage based plans look less attractive than Quota based.

    The reality is that ISP’s cost model is not comparable or transferable to a customer cost model, so artificial models are needed. There’s a few variations on the current model, but the market has worked to remove all but the most competitive (looking) ones.

    Perhaps it is the right time to ask ISP’s to think outside the box, and perhaps the NBN will shake loose some new models.

    • Perhaps it is the right time to ask ISP’s to think outside the box, and perhaps the NBN will shake loose some new models.

      I think the 95th Percentile charging model that Simon (why Is it always Simon floating ideas?) Hackett suggested might be an example of alternative thinking.

      But as others have said, the NBN can only change so much as it is just part of delivering the Internet.

  17. Thinking outside of the box only works when you have a free market pricing model, thats not the case with NBN (and it is the case with ULL/LLS).

    So don’t expect any outside of box type thinking when all ISP’s are forced to adopt the AVC/CVC charge

  18. “If you want to leech, PAY FOR IT!” screamed one commenter. “Plain and simple!!”

    Quote of the century!

  19. Exetel tried the “pay for use” model for both ADSL and NBN (Tassie), but customers are more used to a “pay for a quota which they may not use” – both for internet and mobile phones (ie caps). They therefore had to revert to a standard plan. And what is the point of a limit, when exceeding the limit is almost without penalty? So, unless the public accept a different model, nothing will change. it has been tried, it didn’t work.

  20. So people bagged the hell out of your last article, and your response is…. to do it again? Will wonders never cease? You are just asking for it now.
    No, i didnt bother reading this, i couldnt even get through the last one

  21. I find it fascinating that in this day of modern communication people can’t pick up the phone, check some whirlpool threads, or read the public NBN wholesale cost papers and do some actual research. Making suggestions in an article and then finishing it by saying that you haven’t asked and ISPs and don’t actually know it can be done hardly does credit to your profession. Yes, ISPs bank on average usage that is lower than the quota – if we only billed for actual usage the costs would be higher. The (very public) NBN wholesale cost is $20/Mbps (plus backhaul cost)… if we are talking about shaping for those that are at the top of their quota (and throwing the average out) those people are going to continue to use the data. If I provision 5Mbps of “shaped” bandwidth for each of these users then that bandwidth will cost $100/Mbps (again, PLUS backhaul, plus interconnect and transit costs).. how do you include $100+ of value in a $59 NBN data plan (or, as an ISP – why would you?).. those users throw out the average and cost any profit. When the top 5% of users may use 100% of their quota, you can calculate an average based on that top quota amount – when the shaping allows that to continue at very high rates you can very quickly go out of business (as the leeches will quickly discuss the higher throttle option on the newbie ISP and migrate to them). With ADSL it is different – lower max sync speeds limit damage, and many ISPs have their own high speed fibre connection to the DSLAMs – their backhaul cost may be as low as $1-$3/Mb which is ~10% of what the cost will be under the NBN… if anything, this high interconnect cost increases the cost of traffic for high use individuals under the NBN. Now most of this data mentioned above is in the public domain, it’s hardly an internal ISP secret – so I’m amazed at the opinion, and comments that followed. As a willing participant in the broader ISP community, I’m almost always open for comment… so perhaps next time a little fact checking would go some way towards presenting a sound and valid argument.

    • hi Steven,

      I find your comment insightful, but please keep it polite, as per our comments policy:

      I do speak to the ISPs on a regular basis — please note that I did approach both iiNet and Internode about this line of stories; neither got back to me. TPG won’t comment on virtually anything, and I highly doubt that Telstra or Optus would comment on this either. That leaves what … Primus? Great.



  22. Sure, I’ll bite. :D

    Your subsidisation argument is a good one, but I think it ignores something. “Users are stupid” is a pretty good way to send yourself broke, if you’re going to use it on the sales side, anyway. (On the support side, however, it’s almost necessary.) However, as you say, TPG makes profit and lots of it. To compensate, I would *suggest* (without having any knowledge at all) that perhaps they limit the total amount of bandwidth, such that it doesn’t blow them out of the water. Cross-subsidisation, as you say. However, in times of peak bandwidth usage, you might not be getting those speeds that you paid for. Part and parcel of being with a cut-rate ISP. Internode is different – you pay a premium, and quite rightly expect to get the speeds you pay for all of the time.

    Disclaimer: I am just 6:45am theorising here, I have little to know actual knowledge or evidence to back this up, and I know there’s holes miles wide.

    Side note: I still like the idea of paying by GB, but it needs to be affordable. That was the problem with the old plans (does Telstra still do them? I’m not sure) – no shaping, but the bill shock hurt.

    P.S. Hardly outraged, by the way. Just not overly worried by the idea of being shaped. If I need superfast internet, I’ll buy the datablock. If not, I can check my email with 128/128, and get some sun. :D

  23. While I agree that 128/256 can grind the gears, is 512kbps really so slow as to be unusable? At least for the next couple of years, it should be plenty to get predominantly text-based pages onto screens in a timely manner.

    Whatever number is chosen for shaping is going to be, essentially, arbitrary, but I’m curious as to the reasoning behind the speed(s) you’d prefer to see, Renai. And, if you’re feeling generous, what kind of customer you think is going to be typically needing this kind of backup performance.

      • If I have no self-control, why should I be able to continue to access high-bandwidth video functions at someone else’s expense? Or, honestly, if I am caught with my pants down while downloading too much, is even a 5mbps pipe going to be good enough?

        Assuming a level playing field where the customer is able, but refuses, to buy a data block, make use of their mobile broadband, or upgrade their plan on the spot.

  24. As per my comment with the last article. I wouldn’t even consider anything then unlimited anymore, therefore the whole quota discussion is mute! I use TPG Unlimited and I’d pay more for the NBN version as long as it’s made available.

    Service and customer service for TPG? Wrong! I switched from iinet about a year after having been a satisfied customers with them for a few years. I never looked back though, in terms of service they have the same occasional slow downs same occacional drop-outs, but 99% of the service is good. I am close to the exchange and get 1.6mbs at ANY time of the day, I know I’ve tried. The few times I’ve used customer service, yes they were not local but they were responsive and resolved any issues swiftly, which is what counts.

    And the big thing for me: Just to not have to worry about quota is tremendous relief and the days where I’d have to ration, schedule downloads to cater for peak/ off peak what a huge, huge headache. It’s just not on and apparently also quatos are unheard of overseas (or at least for the majority of countries).

    In fact I use less quote now with unlimited than I used, where as before I was always in the habbit of downloading what I could I know download at my leasure and apparently less than I used to, but it just feels good not being restricted by quota.

  25. I see alot of folk talking about TPG running congested links – I was a long time iiNet customer, not far off 5 years and moved to TPG about half a year ago – if anything my international speeds on TPG feel faster than iiNet – very rarely I have been effected by congestion, but I’d be lying if I said it didn’t happen once in a blue moon.

    Regardless I just ran a few international speedtests, these are pretty typical and I get them all the time. It’s currently 4PM in Sydney and as far as I know the entire country is on school holidays (at least WA is) one could argue this is peak time.

    The slowest is 5.3Mbps. Note I am in Perth so the latency will be a bit inflated – I tested to Tokyo, USA, NZ and most other services nearby – I did a test to Papua New Guinea but I think there is some sort of bottleneck with AU>PNG capacity as it’s always slow regardless of the time.

    In regards to unlimited/shaping I’m on TPG’s unlimited plan and have downloaded 34GB this month – since I installed this usage meter my busiest month has been 160GB with 17GB uploaded – when I first got unlimited I downloaded over 1TB the first and second month but I’m more or less over it now.

Comments are closed.