One of the lessons the femtocell industry might learn from their UMA/WiFi colleagues will be about whether to insist customers take your wireline broadband service with their femtocell. There's a tradeoff here between providing a complete service, one customer care number for all faults and capturing more revenue from the customer, versus putting barriers in the way of early adopters.

[A reminder that UMA/WiFi uses a special UMA capable handset that can connect to the mobile network either over the traditional cellular radio or a WiFi hotspot in your home or other public area. Femtocells differ by allowing any 3G phone to be used, subject to the policy of the network operator and the femtocell owner].

Many mobile network operators (especially in Western Europe) have quietly been buying up complementary networks capable of providing wireline DSL broadband and voice telephony services. Vodafone can offer these in Germany, Italy, Spain and Portugal amongst others; Telefonica O2 in the UK; France Telecom/Orange in the UK. There are also many encumbent telcos who always offered both fixed and wireless services, such as KPN in the Netherlands, France Telecom and Deutche Telecom in Germany.

One of the key commercial drivers for being able to offer the combined triple-play or quad-play services was to be able to capture a greater total share of telecoms spend within each family or business, and reduce the chance of churn because customers become more "sticky" because there is more work and cost involved in switching supplier. Although the overall profit margin per month may be lower, the lifetime value of each subscriber is higher. Dealer subsidies to attract new customers are reduced because fewer customers switch supplier as frequently.

We have seen both mobile and DSL broadband suppliers introduce longer contracts, with high penalties for early termination. It's not unusual for 18 month contracts to be signed for either of these services. So whilst regulators have put in place procedures to easily change supplier, including number portability and migration codes for broadband services, the contractual penalties prevent rapid choice. Few users track their contract expiry dates closely, especially for broadband, and may not be aware of financial implications of switching.

I've documented my experience in trying to buy the Orange UMA service in the UK (which they aren't currently actively marketing).

What can we learn from some current UMA operators' approach?

• Orange UK currently require customers to switch to their wired DSL broadband service and only permit UMA calls when using a WiFi access point on their network.
• T-Mobile US allow users to make WiFi calls from any WiFi connected hotspot, whether on their network or not.

Commercially, it depends on the current contractual situation with the customer's DSL provider:

We've seen a growing trend (certainly in the UK) for fixed network telcos to provide lock-in contracts for their wireline customers. There had always been a minimum of 12 months for standard fixed phone line rental, but this is being extended to cover broadband too - not always through completely transparent marketing methods. Sometimes opt-out bundled package deals are provided at similar prices, but with a contract period included. Other operators have longer contract periods of 18 months or more - effectively adopting the practices of the wireless industry.

Technically, reasons to restrict connection from other broadband networks include

• Better quality of service, because the traffic flows through the same operators network end-to-end and can be marked and prioritised as such. This is particularly true if voice calls can be prioritised by marking them for high priority (using DiffServ which some combined femtocell/DSL modems are already capable of doing).
• Control over the location of femtocells. If they are moved to other countries or switched between properties, then this can partly be detected and restricted by limiting connections only within the operators broadband network.

Several people ask if the quality of the broadband network can be required to meet a set standard and provide good enough QoS, latency etc. to ensure a good customer experience. The internet, even within a controlled broadband DSL service, is still really a best-effort network. Operators do use DiffServ, traffic shaping and other techniques to prioritise and manage the traffic flows between subscribers. This allows them to provide a "better effort" service for premium customers (e.g. businesses paying extra) and low bandwidth/low latency services such as VoIP.

Where end users are provided with a very low cost broadband service that is shared using a contention ratio between 20:1 or even 50:1, its not surprising that some traffic will be delayed or fail to get through. However, the actual bandwidth required for voice traffic is relatively minimal (the standard GSM codec runs at 13kbit/s before overheads).

In some ways, the variable quality of a radio connection between handset and basestation (remember those frequent dropped calls and poor reception in the early days of mobile?) are being swapped for an excellent local radio connection and a variable quality broadband backhaul. For data services, especially where data is offloaded at the femtocell directly, then end users would notice little or no difference from WiFi. For voice, the wide deployment of VoIP services including Skype and others, has proven that these best effort IP services are good enough.

Therefore, I would recommend to femtocell operators that they do not restrict their femtocells to their own broadband networks at launch, but provide some financial enticement to switch later - especially when the current DSL broadband contract tie-in expires.