It was a major regulatory shift that forced BT into considering how best it could replace its 20-year-old OSS customer premises equipment (CPE) product line fulfillment system. It seized the opportunity to increase its right-first-time fulfillment rate in the interests of potentially saving hundreds of millions of dollars a year and greatly improving customer satisfaction.
The key is data driven processes generated outside the OSS, so that changes can be made quickly without recoding or retesting, or destabilizing the OSS itself.
Regulation was the driver
In 2005, after Ofcom's Strategic Telecom Review, BT gave Ofcom an number of undertakings (legally binding assurances) to avoid being referred to the Competition Commission, one of whose options was to split BT up. The main thrust of BT's undertakings was that it would divide its organization up to address multiple markets, ensuring that the divisions' dealings with each other were transparent and not on more favourable terms than with any other service provider.
BT established Openreach in 2006 to oversee the maintenance of its national network and the installation of new infrastructure. It remains part of BT, but was fenced off from the rest of the organization with its own headquarters and staff in the interests of giving all telecom service providers fair and equal access to BT's infrastructure.
Obviously this had big implications for BT's internal systems and processes. For example, it posed a problem for the OSS customer premises equipment (CPE) product line fulfillment system as it was used by both Openreach and BT Global Services, which was no longer allowed. BT decided to move Global Services' CPE product line onto a new system.
The CPE product line generates revenues of ?00 million per year and handles hundreds of orders every day. The product line includes more than 20,000 items, embracing everything from phones, fax machines and cables to Meridian switches for call centers and bank trading rooms. The product line is far from static - last year BT added more than 2,000 new products or modified versions of existing items.
BT opted to build its new OSS around Oracle Communications' Order and Service Management (OSM) solution. This was partly because BT had signed an enterprise-wide licensing deal with Oracle the previous year, so it wouldn't have to pay licensing fees, a big saving. Also, the project team, headed up by Jim Peters, BT Design, chief architect of service fulfillment and workflow platforms, wanted to build an end-to-end stack and the CRM, through which orders are placed, was built on Siebel software. Oracle acquired Siebel in 2005.
The system's main build-out was undertaken in three 90-day cycles and was scheduled to go live this fall. Since Ofcom relaxed the time limit for BT to carry out its functional separation of systems, resources have been switched to address more pressing needs.
Peters insists that much has been gained by the project even before its completion. He explains, "One of BT's core corporate missions is to score 95% and above in customer satisfaction surveys."
In particular, the company is focused on the right-first-time metric, which means the customer gets the right product, in the right place, at the right time and the service works first time.
"The cost of reworking orders and fulfillment for large corporate clients is huge - it typically costs operators hundreds of millions of pounds a year," he said. "If we can get the right-first-time delivery to over 90%, working toward 100%, then the savings are massive and customers are happy. It is the key differentiator between service providers, because the products and services on offer are similar."
Its goal was to take the OSS CPE product line out of the critical path to accommodate changes. BT's internal product owners frequently want to update products and the way they are supported, which involves recoding and retesting the OSS, which makes the cycle much longer than they would like.
The trouble with work-arounds
In the meantime, the operational world doesn't stop so product owners resort to work-arounds such as robotic solutions, spreadsheets and paper-based methods. "The trouble is that comments, phone numbers and date changes fall between the gaps, which has an impact on right-first-time delivery," Peters said.
Which is why BT came up with a solution that could be configured to accommodate the rate of churn among the products without destabilizing the core OSS, he says. "The key was to create a generic fulfillment process whereby processes are defined as data on an system external to the OSS so that they are not destabilized in any way - in this case on the Oracle inventory system, which is then fed into the OSM."
Customers' orders are received via the CRM, then passed to fulfillment to deliver. "Here we check that incompatible products haven't been ordered and then we might break the fulfillment process into 15 steps, say. That might include sending someone out to the customer premises to check they have sufficient power or cooling in place. A lot of the equipment we supply is heavy, so we might need to check the wall it is to be mounted on is strong enough. To install it we might need to order parts from suppliers and need engineers to go out and check the installation before it goes live.
"It might be we find we need to change the sequence of events or make more moves, adds and changes than we initially foresaw."
Instantiating a new workflow
With the existing system, this introduced delays and greatly decreased the chance of right-first-time delivery failing. With the new system, processes are generated by the Oracle inventory system to create a new workflow, for instance, to set parameters to indicate if a task is in jeopardy if it is not completed in four hours and it's meant to take five.
"Or we might drop the jeopardy threshold from four hours to three if we discover the task can be completed more quickly. With the Oracle system we just instantiate a new workflow, so that the next job in uses the latest workflow."
This dynamic approach to order fulfillment that doesn't interfere with the OSS itself has created a great deal of interest within BT, Peters says. "This is a major step forward. Colleagues within other areas are keen to data drive their systems too. The system we've started work on will come into play, but in the meantime we've demonstrated a mechanism where you have an externally defined workflow that means you don't need a hardwired, inflexible OSS."