Analyzing Markets with Network Data
Network hardware spews out Internet operations data fast and furiously, and perhaps in that heap of deep packet inspection (DPI) data… there’s some marketing gold? So is network data that is used for managing traffic and performance useful for marketing insights? There was a time when this was a new question and it made for a creative sales enablement project.
After one “proof of concept” analysis, we said ‘yeah, but probably not by itself. Data always needs context.’
A dozen POCs with Tier 1 Internet service provider data gathered during sales trials later, it was clear a new world was about to dawn. At least for our client. A test service provider found data so useful that it placed service switch in-line in its network. Networks requires redundancy, so it doubled number of switches and the sale became gigantic. They also accelerated deployment for even faster revenues.
And then a little more than a year later, they were acquired.
Grow Sales by Showing Value of Product’s Data
Of all the POCs, let’s focus on the most interesting one – the one that led to the first enormous sales deal. In the quest to know if network data provides priceless marketing insights, one network operator wanted to compare consumers who buy Internet services in two different cities. If there was indeed a big difference – and there is value to the data – they would deploy switches with full redundancy. This means 2 for each location.
Find Marketing Value in Deep Packet Inspection Data
Analyze Users in 2 Cities
Transform Network Data into Consumer Usage Habits
The carrier gave us 2 weeks of network data for 300,000 users on their switches located in two different cities. It was an enormous, albeit straightforward, data transformation exercise.
Traffic data has only a handful of useful variables. There’s a time and date stamp, application variables, and traffic in and out. We transformed it all into a per subscriber view with hundreds of new application usage variables within time of day ranges.
But here is the thing – all apps are different. So getting to the right measures required network technology know-how. For example, streaming is skinny traffic. So number of bytes of streaming traffic alone does not adequately tell the story. Instead, a user’s daily usage of streaming and how it changes are likely better indicators.
Next, we analyzed data to answer these questions:
- Traffic looks the same, but do users also?
- Does everyone use the Internet the same way, including apps and amount?
DPI Data Grows into its Own Business Line
The big reveal: 2 cities had very different app usage patterns. This peek into consumer Internet usage habits proved operations and marketing value. In the end, the telco deployed subscriber management switches everywhere.
And the switch vendor, our client? They scored this big sale, became a market leader in broadband optimization platforms and then got acquired by Arbor Networks.
As for the data and its analysis, it evolved into a real product. Today it thrives as Arbor Insight SP.