Lest We Forget

I have to agree with Russell. The high download charges are a major blockage to mass-market adoption of wap/imode - particularly for off-portal content.

Now, perhaps all-you-can-eat packages are not the answer for everyone, but some will want them.

I suggest that operators bundle a small usage volume (eg 1MB per month) free with atandard monthly fees. Then have options to buy add-ons of 1MB, 3MB, 5MB, 10MB or whatever. The free sample will generate traffic and entice users to spend more, which must be better than them never going near it in the first place.

I do agree that I think it is possible.

But if each customer pays a fixed price per month for all they could want to eat - then your max revenue is capped. Where does the growth come from?

Also: (and this is near term issue) If the most that one can realistically consume on mobile is 4 to 5Mbytes per month (think mass market here not us early adopter geeks!) then why pay (say) 20 pounds per month all you can eat when you could be paying (say) half that much pay as you go?

Just some thoughts.

Jag - wow, thanks for the detailed analysis. On the other hand, all-you-can-eat plans are pretty common with our US cousins, so they think it must be possibel to do without going bust.

Russell

Hi Russell, your point at the end regarding profits is one that could be debated. Wanted to examine this a bit further. Starting from first principles in an attempt to understand the business model:

Everybody knows that Profit (P) equals Revenue (R) minus Cost (C):

P = R - C

Now - in an all-you-can-eat scenario; i.e. where the user has infinite use of the network for a fixed price of €X per month, the Revenue will be directly proportional to the number of customers we have in our system. Let's call the number of customers n. Thus - Revenue is proportional to n - or written as being a function of n:

R = f(n)

Now - Cost will also be proportional to n because there are fixed costs associated with having a user in our system (e.g. a proportion of the radio system, the bandwidth, the order processing, the customer care, the technology etc. etc.) However, because the user will transmit packets to things - and for this particular discussion I am going to assume a worst-case position: that is - because of the infinite usage entitlement, the more users (traffic sources and sinks) that there are in our system, the more packets will be transmitted by each user. Thus - a new order of n is introduced into the Cost proportion:


C = f(n-squared)


Which makes Profit look like this:

P = f(n minus n-squared)

(I am ignoring the effects of costs associated with terminating packets on the operator network boundaries and will assume that these will be settled on a volume basis through the regular settlement models.)

Now, the solution space for Profit P is dependent upon what coefficients of n are applied.

Let's call these constants A and B. Thus:

P = A(n) - B(n-squared)

At first sight, the above looks very alarming - intuitively one would look at the above equation and conclude that over time - as n increases (i.e. as the number of users on the system increases) Profit will tend to zero and eventually become negative! More on this later ...

Now - depending on what unit of measure you want for P - the constant A represents the "€X per month" that all n users will pay for their all-you-can-eat usage entitlement. Thus, A is a very straightforward constant - and is actually something that we would want to FIX rather than DERIVE. Why? Because (like many subscription type service package in adjacent markets - e.g. TV, telephone, DVD rental etc.) there is an upper limit on what users will pay for this kind of service) - i.e. it's something that the market will determine - and in fact, operator choice of A will determine the market that it is addressing! The constant B is little more complicated ...

The "B" factor

In any network business the two biggest contributors to Cost are people and bandwidth. Because there is bandwidth-related cost associated with user acquisition (cell-site and other capital equipment) as well as bandwidth-related costs associated with customer usage (rental and licensing) - the Cost side of the equation is proportional to the second-order of n (i.e. n-squared) and the constant B will represent the unitised cost factor for each user - and will largely be determined by people (i.e. employees) and bandwidth (once-off and recurring) costs of the business. Fortunately - like all good businesses - it is this side of the equation that can (to a greater extent than Revenue) be controlled by the management - and therefore the maths should allow us to DERIVE what the constant B should be in order to generate a certain size of Profit - rather than having it FIXED. (Although I have to admit that to play around with the VARYING of both A and B - as conducting "what-if" analyses are an important part of investment appraisals and management decision-making.

So - back to the point that causes alarm: i.e. surely the Profit will tend to zero and become negative? Not necessarily: only if the following conditions are true:

1. The number of users can grow to infinite proportions.
2. The constant B is truly a constant - and in no way is a function of n.

Condition 1 above can never be true - there is only a finite number of mobile phone users in the markets within which operators play. And besides - the savvy operator would probably want to place an "upper limit" on the number of users it wants to acquire in order to keep the Profit at a healthy and acceptable position above zero.

Condition 2 is interesting because this is alluding to the fact that if B can reduce at the same rate than the number of users n increases - then this will "cancel-out" the n-squared element and transform it into a simple first-order n. We have to ask ourselves if it is truly possible that B can reduce at a rate proportional to n? This is the 64 thousand Euro question! If the mobile data business model is not to be subsidised by something else (like see what's happening in Broadband) then it HAS to. If you accept that the biggest contributors to B are people and bandwidth - then sure - it might be possible to "do more with less" from a people perspective as n increases - but the real gains come from the bandwidth side of B. i.e. for B to reduce over time (in order to extend the lifetime of positive Profit for as long as possible) the cost of bandwidth-related capital and operating expenses must fall - and ideally in line with the increase in the number of users. Is this likely?

Well - there is a side to me that wants to believe that it can: the past 10 years are a demonstrable proof that the cost of technology has fallen dramatically - and continues to do so in the face of increased competition and increased automation etc. And as another example, the explosion of the Internet in late 90s also led to massive falls in the cost of bandwidth. The problem I am told by the network experts in wireless is: the scarcity of spectrum - and the user/unit cost of equipment utilising that spectrum - e.g. base stations, transponders etc. are nowhere near those of equivalents in the fixed world. But use of spectrum evolves - and the equipment that sends signals over it also evolves. My theory is that all you can eat *could* be a sustainable, profitable business when the cost falls as the number of users grows - just like in fixed world. Or it could be is "propped up" by something else - e.g. as part of product/package mix? In the meantime - operators could do all-you-could-possibly-eat instead of true all you can eat.

Just how much could a mobile data user reallistically consume?

I have written about this before in Tom Hume's blog I think - but I reproduce the analysis here again for convenience.

The answer depends on several things: what network you are using, the type of device you are using also, and the context within which you are doing it. Let's try to build a very simple model for typical mobile data usage - just to get us started on this thinking.

Assume that the consumption of mobile Internet happens whilst a) on the move between home and work or travelling somewhere and b) whilst at home or at work when you are not typically near a PC to get on the Internet. Let's assume furthermore that there is approximately half an hour of such time per day that you are using your phone for browsing. Assume that you are doing this an average of say 20 days per month: browsing on your phone. (Personally I think this is very high in mass market terms - but let's go with it for now.) If you are using a PDA type of device - and you are using it to browse full web pages (as in Web'n'Walk) then assume that the average Internet page size is 100k (is that too low?). If you are using a "phone" style of device then assume you are browsing branded portals or i-mode or WAP sites. Assume that the average page size for these is 3kbytes (is that too low?). Now consider the amount of data you could possibly consume in a half-hour given that it takes time to download pages (downoload time) - and it take time to absorb the content on the pages (linger time). On a phone - the download time for pages will be pretty much the same whether on 2G or 3G for 3k size pages - let's say less than 5 seconds if you include the click and navigate times. The linger times will vary greatly - so let's just assume you spend an average of 20 seconds lingering on pages. (That will take into account pages that you just speed through - and pages that you spend a lot of time reading - e.g. news articles etc.). For PDA-style devices - let's assume (for GPRS) that the download time per page is around 60 seconds (bigger pages at around 1.4Kbytes per second) and that for 3G - the download time is 15 seconds ( 6.4Kbytes per sec) - and that the linger time is 3 x that of the phone (richer pages with more detail - and possibility of more time spent navigating a page using scroll bars etc.)

So - for browsing on phones - this very simple/crude model says that you will on average get through 2.4 pages per minute - consuming about 7.2kbytes along the way. Half an hour a day means around 70 pages per day - which means around 1400 pages per month in our model - which is about 4.2 megabytes of data.

For PDAs - our model says that, for GPRS, the page rate per half-hour is about 15 - which means 225 pages per month - which is 22.5 megabytes. On 3G - the page rate goes up to 24 per half hour - which is 480 pages per month - which is 48 megabytes.

Now - it had better be fairly fresh and compelling content in order for anybody to be committed to browse 70 pages per day on a phone 20 days per month and be paying for it. My belief is that this becomes more likely the more content there is available. The open content model is all about having as much content as possible - that is of a quality, choice and freshness that the customer will pay for. More likely than within say closed, walled gardens. (I'm not actually sure that there is enough content in walled gardens to exploit the page rates that can be consumed on mobile phones!) Our model shows that if you had access to great content you would pay for - you would on average consume about 4 megabytes of data per month - which if you look at say the UK price plans for mobile data outside walled gardens - places you firmly at that part of the pricing scale where there is quite a lot of competition in data pricing bundles.

Then you look at the PDA version of the model - and this is one for the those who claim that the over-the-air access to the regular Internet pages will kill WAP/i-mode: Today's pricing for data means that despite the fact that you get through pages much slower this way - you end up consuming between 22 and 48 megabytes of data per month - for which you would have to be paying a lot more money per month. Anyway - the point that I'm really trying to make here is that the PDA-style of browsing the full Internet is still far from being mass-market - not only from a pricing perspective - but also from a user experience perspective. Things like Web'n'Walk might appeal to the likes of me and you - who are comfortable carrying multiple devices - and are comfortable using a stylus and mobile keyboards on devices - that run Windows etc - whilst sitting on the bus or the train - but the majority of people are not there yet. The majority of people are still carrying "normal" handsets - and are just about used to navigating menus on small screens - and are used to typing things on numeric keypads reading and writing messages on small screens - and some maybe are downloading games and ring tones from WAP push type messages.

Is it going top be possible to build a sustainable, profitable business offering mass-market all you can eat for a decent price per month that might be more expensive than they could realistically consume on current data rate plans? I don't know the answer to this question!

Best regards - Jag