Crowdsourcing gives users the center stage with automated network planning

Getting ready to deploy a 5G network? Or expand the existing 4G one? We know that good network planning is fundamental to getting a cost-efficient and high-performance network. And recently, AI and automation have greatly increased our ability to use a wider range of data, with higher granularity in the planning phase.

Gone are the days when operators had to plan their networks based only on the equipment specifications, spectrum availability, financial resources and cell-site availability. Today, operators can use detailed information on the physical environment and its characteristics (e.g., weather patterns, 3D profiles) that affect propagation, simulations of traffic distribution across the footprint to compare different deployment options, and detailed historical data of traffic loads in their networks.

Automation and AI-based learning make it possible to crunch the high volumes of data now available, and use it concurrently satisfy cost, performance, expected traffic load, and environment requirements to optimize network planning.


What is left out? The users

Despite these major improvements in network planning, there is a glaring absence: the user, or, more specifically, the user experience. While related to network performance, user experience provides a separate perspective into how effective networks are in keeping subscribers happy, and in meeting their needs and expectations – a perspective that can refine network planning, leading to networks that more accurately and cost-effectively meet current and expected demand.

The combination of crowdsourcing and automated network planning now offers a remedy. Crowdsourcing makes it possible to collect vast amounts of user-experience data. And the new automated and AI-based platforms make it possible to use this data to complement network planning.  

Types and sources of crowdsourcing

With crowdsourcing, we can collect user-experience data directly from the devices – rather than infer it from network performance. This data can be collected in multiple ways. It can be collected by operators, application providers, or service providers such as umlaut, Tutela or Ookla. Measurements can be active (require user participation) or passive, can include different metrics (e.g., throughput, latency, application-based data) and can have varied levels of location accuracy. These factors determine whether the crowdsourcing data available can be used for network planning and how it can be integrated within the network planning platform.

Adding crowdsourcing to automated network planning

Crowdsourcing data can only be used in a highly automated network planning platform that is capable to analyze a high volume of data and, to be helpful, the data analysis has to be tightly integrated with other data sources. Infovista, in partnership with umlaut, is doing this: it now uses crowdsourcing data in PLANET, its automated network planning and optimization solution. 

For instance, a mobile operator planning for a 5G deployment in a dense area may be able to overlay the new infrastructure in 4G cell sites, or it may need to use new locations, especially if using higher-frequency bands, such as mmWave. In this case, using 4G network data alone makes it difficult to know where within the 4G footprint the demand is highest and, hence, where the ideal location for a new cell site is. The operator may also underestimate demand in areas without coverage or with limited coverage, and not plan for sufficient network expansion there. Reliable crowdsourcing data provides a new layer of analysis and further insight, which helps operators addressing issues like these.

Avoiding the pitfalls of crowdsourcing

The collection and use of crowdsourcing data has great innovative potential, but is still in its early days. Caution should be used to make sure that its contribution to network planning is reliable and trustworthy.

Data quality is the first concern to address. Is the data source reliable? Is the sample size sufficiently large? Does the data adequately cover the footprint? Is the data collected from a sufficient variety of devices? Does it represent the operator’s target subscriber base? Does the over- or under-representation of some user groups create a bias in the data? These are only some of the questions we should ask before integrating crowdsourcing data into network planning.

If data quality is good, the next issue is to determine whether it is the right data to use. Does it have the granularity needed? For instance, does it include specific metrics (e.g., latency)? Does it include data from indoor locations? What is the location accuracy? The responses to these questions determine how the data can be integrated with network performance data from the operator, and how useful it can be in guiding planning.

Privacy and security are also of paramount importance. Was the data collected with user consent and in compliance with local regulation? Does the data collection or use create any security vulnerability?

These are all questions that can be answered, but it is crucial to ask them before introducing crowdsourcing data into network planning – or into network management functions.

Benefits of crowdsourcing

If an operator has access to high-quality crowdsourcing data and uses a planning platform that can use this data, what are the advantages?

Explicitly optimizing a network deployment for user experience is clearly beneficial to the operator. Better user experience leads to more revenues, less churn, and a higher subscriber growth.

But the benefit of using crowdsourcing data in network planning does not come from just improving the user experience. An operator can improve the user experience by overprovisioning the network and this is a frequently used strategy when data about user demand is not sufficiently granular. But it is an expensive solution to the problem.

Crowdsourcing data enables operators to maximize user experience and minimize cost at the same time: it allows operators to channel their financial resources more effectively and to extract a bigger return on their investment.

For instance, an operator may allocate funding to deploy 100 5G cells in a market. By using crowdsourcing data, the operator will still deploy 100 cells, but it will do so in the locations where the subscriber demand is highest. As a result, the cells will carry more traffic. The TCO is the same, but the value extracted from it – the bits/investment, as well as the improvement in subscriber experience – is higher.

Sponsored by

Upcoming events

Open RAN for a Secure Digital Transformation to 5G
January 26, 2022

Panel organized by the Open RAN Policy Coalition, with Monica Paolini, Senza Fili, Jeff Edlund, Hewlett Packard Enterprise, Douglas Gardner, Analog Devices, Inc., and Mark D. Gray, Microsoft Azure

Sparring Partners | Who am I? Identity across digital and physical worlds in the metaverse era
January 27, 2022

Ken Figueredo, More-with-Mobile, and Monica Paolini, Senza Fili, talk about how our digital identity compares to our physical identity, and what that means for future business models and service provider roles

Accelerating your cloud-native strategy
January 31, 2022

Panel at the Fierce Telecom’s Cloud Networking Blitz Week

The road to Wi-Fi 7
February 7, 2022

Panel at the Fierce Wi-Fi Summit on Feburary 7-8, 2022

Streamlining Open RAN Integration
May 25, 2022

Monica Paolini will moderate the panel at Connect(X), May 23-26, 2022, in Denver

In the news

We use cookies to improve your browsing experience. By continuing to use the website you allow Senza Fili to deploy cookies, as detailed in our privacy policy.