MSOs and system operators have been managing network capacity since the first stretch of cable was run in 1948. The demand for the limited bandwidth resources has been on a constant increase for over 70 years now. With video taking a back seat to data in the last decade, the demand accelerated in both quantity and frequency. Throw in a pandemic that shifts the usage patterns at warp speed, and managing capacity becomes a daily, sometimes hourly challenge.
It seems simple enough to look at peak usage vs capacity and create a list of targets to address, but in reality, it is much more complicated. There are multiple factors to include, and the more of them you include, the harder it is to make a good decision. Capital costs, operational expenses, equipment and construction lead times, differing customer use and demand patterns, and much more, all cloud the clear vision operators have into the once seemingly simple game of network segmentation.
In today’s world where data, if not big data, is king, we can now manage all these variables in a way that is intelligent and efficient. We can move from a reactive model with spreadsheets and endless meetings to a predictive and controlled model that removes the pressure from the pressure cooker.
To make the transition to intelligent capacity management, we need to understand all the variables in the equation. Let’s look at them and see how they can be used.
The Value of Customer Experience: Everyone knows a dissatisfied customer shares their experience far more than one that is satisfied. We also know that once a customer is dissatisfied it is difficult, and expensive, to repair the relationship. By incorporating smarter, data-driven decisions into your capacity management program, you do a better job preventing customers from ever getting to the dissatisfied group. Having better insight into how service groups perform over time, and how that performance relates to customer satisfaction allows the operator to be strategically proactive.
The Value of Time: By incorporating intelligence into the capacity management process, we are able to reduce the unpredictable nature the current methods allow. Basing decisions off a single managed metric works well; however, it allows for all the unused metrics to cause unexpected impact. Special requests come in from customer care, government affairs, or an escalation from the “C-Suite,” and these must get worked into the existing plan. The problem is that these are additional work to a schedule that is probably 110% full already. With better upfront decision-making, using actual data from your systems, these challenges are avoided.
The Value of Money: While finding a way to better manage and deploy resources hopefully invokes a positive emotional response from most people, the more important improvement is measured in dollars—dollars that are invisibly spent as a result of missing important data points. While we try to stay in front of the impact curve related to capacity issues, the reality is that not all service groups are created equally. By incorporating the correct data points into the process, and assigning the correct values to this data, we can predictably reduce the amount of dollars spent in operations like customer care, field operations, and customer churn. Additionally, the easily visible dollars spent on construction, engineering operations, licensing, and equipment are able to be applied in a more strategic way, allowing them to have a bigger impact.
Most MSOs are using a single managed metric to determine when to make a change to the service group capacity—peak usage vs. capacity. This metric is probably measured in some enterprise level tool ensuring that the full system view is seen. The results are often imported into an Excel sheet, where something as simple as “sort descending” is used to capture the groups that are approaching the threshold for action. The long history of managing capacity has taught operators what that threshold level is for their operations, based on their ability to react. In reality, the threshold for action may not be tied to the tolerance of usage as much as everyone thinks it is. Challenges like time to cut a PO, shipping, hub design and implementation, and equipment availability are baked into the threshold used. Without assigning a dollar value to these factors, we allow them to play a significant role in managing capacity and the resulting capital and operating expenses.
In an ideal state, the operators are making intelligent, data-driven decisions for capacity management using all the available sources. Call center and field operations statistics, service group demographics, historical use and growth patterns, current engineering designs, vendor lead times, equipment costs and availability, and other customer specific values are incorporated into a data lake. We are then able to apply the power of AI/ML to create a specific model that results in lower operational costs across the board, demonstrated ROI on the decisions being made, allowance for overrides, and AI model training to ensure the decisions are always relevant, and easier overall.
Easy variables: Peak use vs Capacity, 30–60–90 demand growth curves, DOCSIS versioning and transition, call center stats, filed ops stats, hardware costs. Harder variables: Available fiber vs. construction need, Construction lead time, transitions/changes to plant (Digital vs. Analog fiber transport, QAM modulation, Low/mid/high split return, Video compression, CPE costs involved with transition, etc.).
It seems simple and overwhelming all at once, but the rewards are there if you reach for them. Schedule a call with the pureIntegration team to let us help you realize the rewards.