I took part in last week’s DSP Leaders World Forum as a panelist in “Achieving Maximum Operational Efficiency: How Service Providers Can Best Operate at Speed and Scale,” now available on demand. The session was led by Chief Network Officer Iain Milligan of Three UK, and took an honest look at what service providers can do to get out of the doldrums of single-digit revenue growth and high capex.
Keeping EBIDTA margins healthy with existing core telco services won’t be easy, but we did discuss how to get more out of 4G cores using the investments already made toward OSS/BSS integrations, introspection and observability, redundancy, and scale.
It is possible to position yourself for growth opportunities, but you have to do so by thinking about where you want to be in five to 10 years, and then doing all that you can to make customers happy today so that they’re present for you years down the road.
Customer Experience is the ‘North Star’
If you’re a CTO, you’re focus today is, no doubt, on margins – namely, keeping subscribers happy so they don’t churn, and reducing opex associated with what’s costing the most in terms of inbound calls, customer care, payroll, and energy/power costs. You no longer want to chase every bit of red that pops up on your dashboard, and you no longer want to commit resources to programs that won’t impact your customers’ satisfaction and perception of your network and brand. This means customer experience becomes your philosophical “North Star,” guiding where, when, and how you cut costs and trim fat, and where you invest.
You know that capex will hurt your bottomline, so you want to target upgrades to areas in which higher throughputs and lower latencies will translate into more positive customer experiences.
This will, surely, become a growing challenge as traffic from video streaming, work-from-home, cloud and interactive gaming pushes customer expectations higher. You’ll have to get better at resolving issues faster. That will require you optimize and automate, but… only where it makes the most sense to do so.
Automation is an Optimization Challenge
As I discussed before last week’s event in an interview with Telecom TV’s Guy Daniels, automation is an optimization problem. The heavier the optimization, the greater the loss of flexibility. As one audience member at the Forum said, “automation is like concrete that is flexible when you pour it, but then it quickly hardens into something rigid.”
At the same time, automation is one of the few strategic plays you have at your disposal when it comes to driving down opex costs significantly over time; so not doing it is not an option. You need it.
To balance the concrete and challenged opex line, you should focus on the big-ticket items first. Look at your major cost centers, and the major processes in those costs centers, and ask yourself how automation can help cut those down. You’ll likely find that care, operations, customer acquisition, and power consumption are where you’re spending the most. And the good news is that there is plenty to automate there!
But, as mentioned during the DSP event, the number-one ingredient to good automation is good and clean data. And this is where AppQoE shines. If you have a good data set that shows how your customers are feeling when using apps in your network – whether they are happy when doing so, or unhappy (and the why behind it), that’s a really good start for most automation processes. Like I said above, Customer QoE is your North Star.
Automation and AI: Time to Value Matters
When it comes to automation and AI frameworks, you have to accept that you probably won’t get it right the first time, and probably not the second time either. Time to value is much more important than getting the approach right. Prepare to fail fast, and iterate.
Think of Formula 1 racing, where the teams go from “Design-to-Museum” in 12 months, sometimes with hundreds of changes to the car, daily. Many of these changes are discarded quickly, but the car improves substantially during the season. Ask yourself, “How do I get value from AI and automation projects in weeks and months?” as opposed to thinking in terms of "years," and then go from there.
There’s no doubt automation and AI are key for capex control, but when optimizing, you have to continually assess the key variables that will impact capex and planning for upgrades, such as:
- Bandwidth and the impact on resource utilization
- Application and traffic volumes
- Traffic types and growth
- Subscriber types and growth
- Congested nodes/sites of the network
- Near-to-congested nodes/sites
Insight into each of the above variables is what will set the stage for automated closed-loop approaches to capacity planning and network optimization. Those approaches will have to be driven by customer quality of experience: what your customers feel,when engaging with applications and consuming and creating content on your network.
At Sandvine, we call this “App QoE,” and it is the key factor, not “tonnage,” that should drive network upgrade strategies.
To improve App QoE, you have to take the time – over years – to build out data sources, data science teams, machine learning, NPS surveys and correlations. As I've explained here, this is not a one-and-done deal that takes three months to do. This is a long-term strategy with tactical projects along the way.
To see what we can to help with our short-term and long-term goals, schedule a demo and check out our recent press release about App QoE with Telenet.
Also, check out our other resources:
DSP Leaders World Forum panel discussion “Achieving Maximum Operational Efficiency: How Service Providers Can Best Operate at Speed and Scale”
TelecomTV interview, “Understand Your Network Data to Better Serve Your Customers”
AppLogic How to Reduce Costs and Improve Customer Satisfaction