« Credit Crunch - Silver Lining for Telcos? (Part 2) | Main | Ring! Ring! Hot News, 6th October 2008 »

Why 3G ­Embedded Notebook Forecasts are Overhyped

We’re delighted to be working more closely with Dean Bubley, one of the most insightful analysts on wireless technology and author of the superbly acerbic Disruptive Analysis blog. Dean will be helping us to create more tangible roadmaps to the two-sided business model.

After our upbeat post on embedded broadband, we asked Dean to share some of his detailed analysis and give us a ‘reality check’:

The last few weeks have seen much fanfare about notebooks shipping with built-in 3G modules. Vodafone announced it was selling Dell’s new Mini 9 netbook, while T-Mobile is working with LG, Acer and Asus on embedded notebooks of various types. The GSMA has just announced its “Mobile Broadband” certification scheme and sticker for PCs, backed by $1bn of marketing. It hopes this will emulate the past success of the WiFi Alliance and Intel’s Centrino badging in driving the “attach rate” of embedded connectivity. In the background, various silicon and module providers (notably Qualcomm and Ericsson) have been loudly evangelising their products.

Disruptive Analysis has been working on a new report, due for publication soon, on Mobile Broadband Computing, which examines the various options for connecting PCs and new “MIDs” (Mobile Internet Devices). One early finding of the research is that embedded-3G (and, for that matter, embedded-WiMAX) notebooks will not follow the rapid growth trajectory of WiFi. Instead, the overall proportion of PCs attached to mobile networks will grow relatively slowly, and those that do will use a mix of dongles, embedded modules and other options.

Really, there are two separate questions here:

What percentage of notebook buyers really want or value mobile broadband at all? This defines the addressable market as a subset of overall laptop shipments.

For those that want mobile broadband, is an embedded module better than the alternatives?

The first point to make is that a sizeable proportion of notebook users have little need for mobile broadband in any form, or little ability to justify it on cost grounds:

23% of notebooks are sold to emerging markets - a figure that will likely rise to 30%+ over the next few years. At present, there’s no 3G in China or India yet, and it will still be immature in 2010. There is little point putting in expensive, unused, functionality into PCs being sold to cost-sensitive markets. 60% of notebooks are sold into businesses. Corporations are still only slowly adopting mobile broadband, and the majority of employees still do not warrant the extra cost. Senior executives, and frequent travellers who would otherwise rack up large ad-hoc WiFi bills are the key targets - not the average worker who occasionally takes his PC home at the weekend.

A sizeable fraction of notebook buyers purchase large “desktop replacement” PCs, and leave them in a static location at home or in the office. These have little need for mobile broadband at all, especially in regions with poor indoor 3G coverage.

Overall, the realistic overall “addressable market” for mobile broadband remains well below 50% of all notebook purchasers. Within this segment, it is then worth considering the strengths of embedded options, against alternatives like dongles, or 3G handsets used as “tethers”. Disruptive Analysis believes that the comparison of the uptake of embedded 3G, with that of WiFi from 2003, does not bear close scrutiny, despite the industry rhetoric that the two are analogous:

When launched, Centrino included considerably more than just WiFi - it also significantly improved battery life and heat dissipation, enabling thinner and lighter laptops. There is no equivalent “side benefit” from 3G modules. The price premium of integrated WiFi capability essentially fell to near zero in 2003. Conversely, integrated 3G in notebooks still commands premiums of $100+, and even though new modules are getting cheaper, Disruptive Analysis believes it will take several more years to fall to $20 or below, particularly for the most up-to-date chips.

WiFi adoption was catalysed by a variety of access models - especially private WiFi in homes and offices and one-off public hotspot access. While 3G offers greater general coverage, it not useful for home Internet-sharing, or use as a LAN extension inside corporate offices. The main competing form-factor for embedded WiFi was PCMCIA cards. These were expensive, fragile and required complex driver installation. Today, USB dongles are cheaper, easier to use/install/sell and relatively robust. Phones acting as 3G “tethers” are also becoming more practical.

Put simply - embedded 3G in 2008 does not add as much to external 3G, as embedded WiFi added to external WiFi in 2003. It is also critical to recognise that only around 10% of WiFi notebook owners generate services revenues through hotspot use.


The industry’s motivations for wanting 3G notebooks are understandable - operators and the GSMA want to get further momentum for HSPA ahead of the imminent launch of WiMAX-enabled PCs. They feel (probably correctly) that more users will be tempted to sign up for services if the capability is in the PC when it leaves the shop. They’re also very keen on charging €35 a month for a full notebook, rather than just €15 for a 3G dongle on the same data plan - especially if their auditors allow them to book it all as “data services” revenue, rather than equipment resale, or if the customer continues to pay at the same rate after the end of the contract.

The PC OEMs are looking hungrily at the operators as new distribution channels, with billions in the pot for subsidies - and the possibility of moving PC users to a 2-year upgrade cycle. And the silicon vendors want to sell high-end chips - and stimulate demand for their network-infrastructure products as well.

For end users, the benefits of embedded modules vs. dongles or tethered 3G phones are less clear-cut. Potentially, radio performance and battery life can be better. Maybe there will be some new pre-installed PC-based operator applications that are seen as valuable, perhaps even based around forthcoming Telco 2.0 business models, but these are still in their infancy. Most of the other advantages are generic to all forms of mobile broadband - even the option of subsidised or “free” PCs is possible by bundling the notebook with a dongle.

And these benefits are set against downsides. Some operators’ data plans for embedded notebooks are more expensive and use slower modems than their own dongle offerings. Built-in 3G notebooks risk users being tied to specific operators on long contracts, perhaps with onerous “fair usage” policies, difficult-to-remove operator-custom software, or even application blocking. Where they are sold via operator channels, there are huge outstanding questions about the quality of technical support. Roaming is expensive, and switching SIMs to those of a different provider while travelling may cause glitches.

Nevertheless, there will certainly be a good number of consumers tempted by embedded notebooks and netbooks, especially as the price delta versus “unconnected” ones falls to perhaps $30 or less. One possible factor which could drive greater uptake is the availability of innovative business models and payment methods - something already seen in PCs sold with fixed broadband via Microsoft’s FlexGo proposition. On the other hand, any shifts in perceived reliability of HSPA as networks get more congested - or embedded notebooks start to be used in homes with poor coverage - could put a brake on the market.

If all goes well, consumer notebooks sold in developed markets could enjoy WWAN attach rates of 15% in 2009 - especially lower-end netbooks. Given innovation and acceptable network performance, this could potentially grow to 50% by 2013. However - not all of these will necessarily be “activated”, or if they are, use could be occasional rather than a typical monthly subscription.

For enterprise customers, the total cost of ownership (TCO) is still too prohibitive to offer mobile broadband as a default option for all staff. And using external modems enables them to maintain a single low-cost notebook “build”, but give broadband selectively to those that justify the extra spend. They can also easily switch to other operators if they choose, or even use WiMAX if it becomes available. As an estimate, for business laptops overall, we will maybe get a 7% attach-rate of embedded 3G in 2009, perhaps increasing to 25% in 2011, and more beyond that point.

Overall, the industry needs to see beyond the PR blizzard about embedded 3G in notebooks. Yes, it will be an important trend. But it won’t happen ubiquitously, and it won’t happen overnight. And there is also an open question about what proportion of embedded 3G will ever actually get activated and used - or generate meaningful mobile broadband services revenue.

[Ed - you can meet Dean at the Telco 2.0 event on 4-5 Nov, London.]

To share this article easily, please click:


I disagree with some of the analysis of Dean and have been voicing this on his blog as well.

For one, his analysis is flawed and overly simplistic and lacks context.

For example:
Dean states:"23% of notebooks are sold to emerging markets - a figure that will likely rise to 30%+ over the next few years. ... and . 60% of notebooks are sold into businesses."

What he does not say is that sales projections for notebooks in 2009 is 139 million laptops!

60% = 83.4 million in 2009
23% = 31.97 million in 2009

He then goes on to state:"consumer notebooks sold in developed markets could enjoy WWAN attach rates of 15% in 2009"

Calculate this:
100-23=77% Developed market = 107 million.
107 * 15% = 16 million units in 2009!
According to Dean's own analysis, this is not a small amount. In fact, I'd bet that this will already create more connections the all of Starbucks hotspots combined.

Don't get me wrong. Dean has some very good points. But also some bad assumptions:

1. Operators do not really want to sell notebooks. It is just what they are used to doing (like selling Mobile Phones). They need time (2009) to realize that the best channel is the OEM. Dean keeps focusing one subsidy. its not a model or strategy. Its MNO trying stuff to see what works.

2. Dean makes a mistake in assuming that corporates do not want to provide embedded modules to people because of cost. The cost of purchase is not the factor in a corporate environment (maybe in SME).
In their case its more about the ROI. The same logic that allows them to give blackberries, allows for providing embedded modules. The (illusion) assumption of increased productivity. The less cost of helping someone setup a Wifi Connection at home, the enhanced security features one can use with embedded modules, like the Gobi built in GPS add-on. Less loss of dongles, etc...
We have certain corporate customers that even buy unlocked dongles just for the switching ability with Sims. So purchase cost is not a big factor, as long as there is tangible ROI.

3. Dean mentions "attach rate". First of, this term is incorrect. "Attach rate" in embedded terms is the "industry" term for people actually switching on the radio (ie. getting a subscription) on an embedded notebook. And yes, 7% is the current number. The actual shipped units is not a public number. And I think Dean talks about "shipped units". Plus Dean seems to think that the RF will be better. so far, all tests with 3G notebooks have shown poorer coverage, because the antenna is in the screen next to all the other antenna's. This has been widely discussed over the past 18 months.

Dean also states that one negative point is difficulty of churning because of (drivers) conflicting issue. This issue has been worked on for the last 2 years. First, Gobi addresses this issues as the firmware is flasheable per operator, the Gobi SDK is independent and the most importat; The OEM's are launching with Operator independent connection managers. So this statement is completely incorrect!

Lastly, the "3G inside" sticker has been on the wishlist and debate list for 2 years already. Insiders know this. And frankly nobody (should) care(s) about the sticker. Its ego massaging for the GSM boys. Nor should it matter. The trick that will make or break the adoption curve will be the distribution and price model.

In short, I am hoping that his analysis document, soon to come out, will be better then this topic.

I think it is important to note that operators have not shown enough innovation on pricing for mobile broadband. My current laptop came with a Verizon card built in for only a marginal extra cost. I have frequently used their daily service ($15) when traveling (personal & business). It is similar to slightly more expensive than a wifi hotspot charge but I get the added value of using it for 24 hours rather than just while I'm drinking my coffee.

I think that with more marketing and more daily/hourly pricing options even the average consumer would utilize built in mobile broadband when out of wifi coverage. The use cases for this are numerous including on trains, at the airport, in the car, etc. By building it in to the laptop and innovation pricing models, operators stand to gain data revenue that might not otherwise realize because non-business travelers are less willing to buy extra hardware.

The Blackberries are the real killer of 3G embedded laptops. BB does the job well for email - by far the most desired Internet app. It wins on convenience too - no need to open the laptop to get my email - that is for business. For consumer - add multimedia to the email and this gets things even worst - with the prohibitive caps and narrowband experience of mobile Internet the thing is hardly going to fly for personal use. Moreover - WiFi was a go to a large extent because of low cost WiFi routers, perhaps if there ever is viable case for femtocell...

Overall I am not very excited about 3G broadband.

"Positive user experiences and heavy promotion by leading carriers and service providers will help fuel the demand for affordable and reliable home and small-to-medium business VoIP solutions"

to read more: http://www.communicationstoday.co.in/oct2007/next-wave-in-communication-819-41.html


As frequent comments on my blog have indicated, you have a different opinion on this issue to me.

You also have a specific vested interest in promoting the concept, as I suspect it benefits your company's connection manager products.

You have misrepresented several of my comments. Most importantly, your 107x15% calculation ignores my reference that the possible 15% attach-rate specifically relates to CONSUMER notebooks. ie Not business.

[FYI the term "attach rate" has different meanings to different people. In the mobile device industry, it is used to refer to the installation of a given feature, not usage - for example the "attach rate" of GPS or Bluetooth in phones refers to hardware in the device. Intel has frequently referred to "attach rate" for the inclusion of WiFi in notebooks. Some service providers use it to refer to activation rate. I'll highlight the ambiguity in the report]

I've been consistently skeptical of embedded-3G notebooks since April 2006, and thus far I've been proved absolutely right. I recognise that some of the early difficulties have now been fixed, and a variety of new business models and technologies have become available. However, I still don't believe that this is a megatrend, but a continued slow evolution that is in no way comparable to the adoption of WiFi in notebooks.


(Sidenote: This is Year 3 after introduction of the first embedded-3G laptop. The attach rate is much lower than Year 3 after intro of WiFi notebooks - and that was in the depths of the 2002 bust)

There is already data showing that 3G access to the Internet while "on the go" is catching or passing paid WiFi hotspots. This is no surprise, whether it is dongles or embedded 3G. I have often written that hotspot providers have squandered a 4 year lead in the market through ridiculous pricing strategies. But although hotspot use tends to get all the press focus, it is one of the least important use cases of WiFi. Use for shared Internet access at home, or ability to move about workplaces are far more important.

You mistakenly suggest that I focus solely on subsidy. That is false - read the piece again. I am quite aware that a variety of pay-as-you-go and other tariff plans are likely, as well as (possibly) Telco 2.0-style ones sponsored by third-parties like VoIP providers or advertisers.

On enterprise, I am referring to TCO, total cost of ownership. I am not alone in expecting enterprises to remain highly selective in which employees they give mobile broadband to - Gartner has said much the same recently, and I have heard the same from multiple other sources. As you say, the same logic as BlackBerries - which are typically only given to a small fraction of staff in typical organisations, and not even all mobile users. When it is proven that TCO is lower with embedded vs. dongles I'd expect purchasers to start switching, but that will take time. And as the costs come down, companies will also look to extend the reach of employees entitled to mobile broadband - just the same as with push email / BlackBerry.

The RF point is interesting - the companies I've spoken to have made a strong argument that the embedded antenna is one of the major advantages. If it isn't (and I don't have a personal testing lab) then that removes one of the major perceived benefits. I hadn't detected the mismatch of expectation vs reality - although I have queried vendors about multi-band support (eg HSPA at 900MHz, 2.6GHz etc)

Gobi's ability to blend HSPA+EVDO is worthwhile for some frequent intercontinental travellers. It is less relevant to most consumers. And I also expect to see dual-standard dongles. I refer to it in the report.

I also know that the driver & CM issues have been worked on, but I expect there to remain a significant number of exceptions - for example if a user plugs a dongle into a 3G-embedded PC for some reason, rather than just the SIM. (eg an LTE or WiMAX dongle when the module is HSPA). There will be other scenarios too potentially.

The extra cost issue is a real one, where the embedded module is not an option (eg bought instore rather than online). As I suggested in a previous post, I expect "embeddable" PCs with a slot to become very common.



Thanks for the response.
As odd as this sounds, I think we are not that different of opinion (in a lot of cases).

I agree (and have never disagreed) that embedded notebooks will not have the same curve as WiFi enabled notebooks. Anyone who thought 3G embedded was going to be mainstream from the beginning, starting in 2006 was drinking some spiced cool-aid. :-)

But I believe that sometimes things need to be put into the right context and meaning.

Suggesting I have something to gain commercially with my comments is a shame.
It demeans the time I take to give my expert opinion about a field I know something about. We may disagree, fine. No problem. But please don't think I have a hidden agenda. I write as Edsard, the person. Not the COO of Diginext.
Go back to all my posts and see if I plugged our company or services anywhere.

In fact, this process of embedded 3G, will in the end, not be good for Connection Manager companies at all!
One of the datapoints that helped WiFi adoption grow was also the embedded WiFi software in the XP OS. And frankly, I know that "Mobile Broadband" will also be embedded in the OS. Regardless of Diginext, this will be good for the market. That process is also ongoing for the last 18 months.

Advantages of being in the OS:
1. All datadevices will need to conform to one common API (Not happening in Dongles).
2. No need for separate Connection Managers from anyone, whether it be OEM or MNO. Thus less conflicts.
3. This will also solve the embedded vs. dongle conflict scenario you mentioned
4. Increased stability and less interoperability issues.

Already today many MNO's take the standard CM from the device vendor. So embedding the CM in the Notebook just means less customers for CM companies.
So my personal opinion is not related to the company I work for.

If I made a mistake with the math I am sorry and please correct me. But I did the best with the numbers in your article. I like numbers, not percentages, personally.

And it seems to me that often, in the article, percentages are quotes as "matter of factly" and not opinions, which they are:

For example you state:
There is no equivalent "side benefit" from 3G modules"

It should have said:"According to Disruptive Wireless, There is no equivalent "side benefit" from 3G modules". Or as far as we know...

The embedded GPS and Data does not have advantages?
What about;
Tracking and Tracing?
Theft Recovery?
Lower insurance premiums?
Can you do that with a dongle?

Anyway, All I was trying to say is that in the case of embedded 3G it's not a question of "if" but "when". And I had the feeling that you are/were at the point of "if" it happens.

That's all.

- Edsard

Dean have to say that this might be one of the reports that ends up on the mistake spike rather than you being able to say told you so post.

Looking at the way people are consuming laptops they are something that a user has for a shorter period. The people I work with are now capable of replacing their machines on a 9-12 month cycle, we are highly mobile and so I admit stress test them. However the same seems to be true with our children in secondary education.

This increasing upgrade cycle means that the manufactures can adopt new technology faster. The Sony Vaio range is being updated 3-4 times a year at the moment. This means that in the last year my laptop has moved from a plug in module to embedded mobile broadband. Over the next 12 months I am confident that we will see more of the Sony range come with Embedded chips. Dependent on the price point we will see either Qualcomm or Ericsson chips.

The inflection point comes when the Laptop maker does not see the need to do a deal with the Mobile Network but rather has the APIs for all available networks.

Post a comment

(To prevent spam, all comments need to be approved by the Telco 2.0 team before appearing. Thanks for waiting.)

Telco 2.0 Strategy Report Out Now: Telco Strategy in the Cloud

Subscribe to this blog

To get blog posts delivered to your inbox, enter your email address:

How we respect your privacy

Subscribe via RSS

Telco 2.0™ Email Newsletter

The free Telco 2.0™ newsletter is published every second week. To subscribe, enter your email address:

Telco 2.0™ is produced by: