Keeping BT's tech agile

Can you tell us more about the "One IT" program and how it is instrumental in BT's overall IT transformation?
One IT is the name of the organization that's headed by the group CIO. On the management board are each of the CIOs and the person responsible for the data centers and operation management of the entire IT estate. Through One IT, we have oversight of the entire investment that we make in IT every year, which is about ₤1.5 billion (US$2.9 billion). All of the IT resources are actually located in the One IT resource pool. The CIOs individually have a number of personnel--whom we term "business engagement" folk--working with us to build very tight linkages between the lines of business.

We then organize the bulk of our people into two structures--the first is around what we term customer experience programs, because we believe that absolutely key to the success of BT going forward, is a ruthless focus on improving the customer experience in dealing with each segment of BT. Then we describe within customer experience programs three key experiences that we focus on: concept-to-market (how we bring new products to market), lead-to-cash (a customer's journey from enquiring about a product to purchasing it) and trouble-to-resolve (a customer's experience of having an issue to getting it resolved). We're very focused on having structures around customer experience because in a fast commoditizing marketplace, customer experience will become essential for companies like ourselves.

The other structure is around platform programs. Historically, among large corporations, IT has been designed, built, and deployed for hundreds of business units in a company. That has given rise to problems such as proliferation of technologies, inconsistency of service, and the old chestnut of stove-pipe systems based on segments or products. This has made it difficult to sell converged and bundled products. Our strategy is to define an IT architecture for the BT group that has 14 major IT platforms, such as billing and service management. We encapsulate those platforms with usable capabilities, which are exposed via Web services.

A recent survey by U.K.-based Which magazine indicated that just 23 per cent of BT's customers were very satisfied with the overall level of service they received. Customers also said BT's bills were the most difficult to understand. Do you have any thoughts on that?
The numbers indicate there's much optimization of customer experience that's yet to be done. The companies that do that very effectively will differentiate themselves from the rest of the pack. The evidence reinforces my thesis that you will win in the marketplace not just by having great products and being early to market, but by offering a great customer experience.

BT has a scouting program to acquire technologies that will be used on its 21st Century Network (21CN). Is there a similar program for IT talent?
We do source our technology globally now. In 21CN, you'll find American, European and Asian companies that we work with. It's truly a globally sourced piece of innovation, drawing on the best companies and talent across the globe.

We do source IT talent globally. If you look at our hires in the last 12 months, you will see that we are recruiting from the best U.K. universities, but you will also find people from Europe, Asia and America. If you want the best talent, you shop globally, not just in your own backyard. There's global competition for top talent and you have to be out where the talent is emerging. We're also looking to new and untapped sources of talent, such as Russia.

India, the world's outsourcing center, is well-known for its highly proficient IT workforce. Do you see India's position being overshadowed by other countries such as China?
Our relationship with India is extremely longstanding. I've been sourcing for IT talent from India for 12 years. We have a very large joint venture in India called Tech Mahindra. As a source of IT talent, India is up there in the top five. They have a tremendous education system with a big focus on technical subjects, with a significant number of good universities that generate large numbers of graduates every year.

I find the Indian IT guys very well-educated, hardworking and willing to learn. Often they graduate with IT skills, and the development for them is in understanding the businesses they are serving. India most certainly is not going to disappear as a major source of IT services in future.

They do face challenges in supply, however. The number of IT companies has surged over the last few years. Inside India, there is a very ferocious battle for talent, and that's reflected in significant wage inflation of 20 percent.

We're also seeing lots of churn. You can build teams fast in India because there is an abundance of skilled labor, but can you retain those teams? One of the things that you need to do to deliver and sustain very large IT platforms is you've got to maintain consistency in personnel for a reasonable period of time.

As prices go up and supply problems become more manifest in India, I'm sure other markets for talented IT people like China, Russia and Eastern Europe will emerge. But I think not only will you see companies coming out of those new locations for IT services, you'll also see Indian companies establishing bases in countries such as China.

What have been some of the challenges, as far as the work on 21CN is concerned from an IT perspective?
The 21CN program is hugely ambitious, because unlike other service providers, our goal is not to build another overlaying network based on IP (Internet Protocol) and MPLS (Multiprotocol Label Switching). Our intention is to replace our current networks with one converged core network that's truly multi-service. We will carry voice, data and other services on one network.

One of the key challenges has been collaborative work with the network partners, because many of them are building elements of the 21CN network. For example, we work very closely with Hewlett-Packard (HP) to build key domain management systems, and we have to facilitate collaboration between HP and Fujitsu, which is pioneering in the network's access domain. The challenge is in getting our big IT vendors to collaborate with network vendors in a new space, where we are establishing standards as we go.

Another challenge in the IT space is we're trying to do two things at the same time. One is to build IT to support a new [21CN] network, while changing our legacy systems for our new strategic IT platforms. That's immensely challenging, but extremely exciting. It's a bit like changing the engines on the plane while you're trying to fly along.

The other big challenge is in transfer engineering, as we term it. While you can build a new set of network and service management systems for a new network, the parallel challenge is how you migrate 30 million customers and their data from old systems onto a brand new network and systems without a glitch. That's a mammoth task.

Also, as we explore the customer data on existing systems, we've to tackle data quality issues during the migration process. We not only have to transfer data from old to new systems, we also have to go through the rigorous process of cleansing that data, while eliminating data errors that were built up--not just over years but decades--on some of these older systems.

The completion date for 21CN has slipped two years later, to 2011. Critics have hit out at the delay, especially those who have been looking forward to a nationwide ADSL2+ network. Any thoughts on that?
Well, let me assure you that the level of effort and focus we've applied to putting in place this network, as well as the new systems and processes that will support it, is unparalleled. Nobody is doing a network on this scale anywhere in the world. We are at the leading edge with our vendors, so we do have to collaborate to overcome technical difficulties as they arise. This is pioneering work, and we will get there because we are fantastically focused. We've built an incredibly competent team and that team spans not just the people at BT, but also the vendors who work with us in the network and IT space. And the job will get done in the shortest feasible time

We will learn an enormous amount in the next few months as the first customers transfer to the new network. We will get feedback from our customers, service providers who utilize the underlying network. All of that feedback will be critical in helping us hone a realistic plan for deployment across the whole country. We will get there as far as it is humanly possible. And the result will be the world's first nationally deployed multi-service IP network.

BT's group CIO Al-Noor Ramji has famously said that Google is your biggest threat. Do you think that's way exaggerated? Google doesn't seem to be interested in data pipes after all.
There are a number of aspects to Al-Noor's comments. One is the speed at which Google delivers products to the market. One of the points he's making is Google has a fantastic model for delivering very rapidly, new services to their customers. They have a very clever way of cultivating an ecosystem of application developers around the globe, who can work with Google's APIs, create services and take them to market in a beta sense.

Google has a fantastic model for the introduction of new services: If customers are drawn by a new service, then more work goes into that service, which will be productized and made more reliable and scaleable. I think what Al-Noor is saying is that companies like BT have to deliver products to market in a more agile way, and Google is a great example of that, using resources not just within the company, but creating an environment that encourages an ecosystem of other developers to bring their products to the market.

Al-Noor is encouraging us in BT to think about how we can get more companies and developers involved in creating BT services, or services they sell that sit on BT's infrastructure. He is using Google's model to challenge us to figure how to get our products and services to market more quickly. He's forcing us to think about doing away with the traditional telco idea that everything has to be engineered to scale, at launch. He's forcing us to think if there's a more agile approach about going fast and creating beta style quality [products] at launch, testing with the market where early adopters go for it before doing a second phase of making things more robust and scaleable.

Has there been any progress since then?
We've made huge strides. We've had a significant adoption of agile techniques. Everything in IT is now oriented around a 90-day delivery cycle. We bring together IT, product, service, finance guys as well as our IT suppliers and customers into a 'hothouse' environment, where we tackle issues like how to bring a product to market. We're also focused on co-locating development resources, and iterative development, where in the later phases of the 90-day cycle, we create daily builds and conduct continuous integration and testing.

We've a whole suite of tools now around the agile methodology, which encompasses not just IT delivery but also in how we work with our business colleagues and customers to help define the 90-day deliverable, through a 'hothouse' process. Hothouses always generate prototypes, so customers and business partners can see what they will get in a product in the end. This tackles some of the traditional IT issues at big corporations, where it can take 18 months to make a deliverable, and then discovering that the world has moved on, and the deliverable is no longer relevant.

You've implemented CRM systems for BT Retail. In the past, too many CRM implementations were led by technology rather than processes. How do you avoid such mistakes?
There was a period when CRM technologies were kind of the flavor of the month. A number of tests exist for whether a company's CRM is in good health. One is, it must be focused on the customer, not to be led by technologists. Technology is one plank of it. If you are focused on the customer, you're focused on excellent processes and you're focused on people, because ultimately a lot of interaction with customers is through people. You need to make sure people are appropriately skilled and behave appropriately in particular circumstances.

Another plank is product. Having a CRM strategy without a focus on having the right product at the right time, price point and customer experience would be an invalid strategy. As far as CRM is concerned, the trick is to have product, process, people and technology--all lined up in a single customer-focused strategy. Mistakes have been made in the past across the industry in leveraging some of the CRM packages. A bad CRM strategy would be one which is very highly-configured. If you go down that route, you end up with high cost of maintenance and a really tough job in making changes. You also end up with a major problem in taking new releases of the underlying packages. We are very rigorous about leveraging the out-of-the-box capability of CRM packages, and we're very wary of significant levels of customization.

What are the IT challenges facing telcos today, especially in an increasingly converged industry?
For a company like BT, IT is moving from a support function to a revenue-generation platform of the future. An example would be IPTV, which we are launching in the UK. That platform is an IT platform. It's a content management and distribution platform. It has all of the controls that are needed to guarantee bandwidth and quality of service on the underlying broadband network. All of that is delivered though IT. The big challenge for IT guys in telcos is to understand that they are moving from the periphery of the company where they just provide billing and accounting systems, to center stage. They are moving to the very heart of the telco, because they are building the revenue-generating platform that represents the future of telcos.