Big Data Analytics Improves Planning for Broadband Projects
Building out a new fixed broadband network is a massively capital-intensive undertaking, and one that operators' senior management and investors do not take lightly. Go/No-Go decisions rely on key financial metrics -- Return on Invested Capital (ROIC), Time to Break-Even, Time to Positive Cash Flow, Debt Service Coverage Ratio. These have numerous dependencies, especially:
- customer density (potential customers per km)
- expected take rate (percentage of customers passed expected to pay for the service)
- expected ramp rate (monthly increase in take rate, after the service is open for sale)
- average revenue per account (ARPA)
- capital expenditure (capex) per customer passed
- capex per new customer acquisition
- time to first revenue
These factors vary widely by neighborhood geography, demography and field conditions.
For practical reasons, broadband network rollouts are typically phased. Preliminary network designs place local convergence points at optimal locations, dividing the project area into serving areas. "Fiber zones” are units of geography for project scheduling and marketing, each consisting of a small number of serving areas. For each fiber zone, financial analysis and business constraints determine its priority in the project schedule, or, perhaps, to serve it with an alternative technology or not at all.
Google Fiber pioneered use of fiber zones as a tool for network planning and marketing. Its methodology uses customer surveys, pre-subscription and grass-roots marketing to forecast take rate and ramp rate. While a great improvement over ad-hoc planning and marketing, it misses the many customers who do not respond to the pre-subscription campaign, despite being interested in the service. More important, it also does not forecast ARPA.
Big data analytics can predict each potential customer's interest in specific services and tiers, price sensitivity, budget, propensity to switch service providers, susceptibility to special offers -- and thus their value as a customer. Aggregated, these consumer behaviors can estimate take rate, ramp rate, churn, and ARPA for a fiber zone, as inputs to financial models. The data is subsequently used for marketing and customer care throughout the service lifecycle.
Customer analytics for planning isn't just for cherry picking profitable neighborhoods. Revenue from early customers of a project offsets capex for subsequent construction, improving the project's cash-flow profile. For debt-funded projects, it covers early interest payments and reduces total borrowing. By maximizing early revenue, analytics-based planning can make a difference in the financial viability of a broadband project.
This blog is sponsored by Huawei.
Dan Grossman, Contributing Analyst, Heavy Reading