Last week, I released a new plan finder tool. Users accessing the tool can answer a few questions about how they use their phones, how budget-sensitive they are, and where they live. They’ll then be matched with a few carriers and plans that are likely to be well-suited for their needs.
A few other companies have released their own plan finder tools. These tools generally function by assuming the wireless industry is simpler and more commoditized than it is. For example, WhistleOut’s tool appears to assume that cell phone plans have only five features:
- A host network
- An allotment of data
- An allotment of minutes
- An allotment of texts
- A price
The allotments are all assumed to take fixed, numerical values. Plans’ prices are also assumed to take simple, fixed values. The host network is simply one of five options (Verizon, AT&T, T-Mobile, Sprint, or U.S. Cellular). Making these assumptions allows many carriers’ plans to be compared, sorted, and filtered with basic math and logic. Unfortunately, the assumptions sweep a lot of important nuances under the rug. For example:
- Carriers may throttle data speeds or ignore data use from certain applications. Complicated data policies can’t be captured when assuming that plans have simple, fixed data allotments.
- Pricing may not be fixed. E.g., Mint Mobile has one price for subscribers that purchase 3 months of service upfront and another price for those who purchase 12 months of service upfront.
- Two services that use the same host network could have different levels of priority during congestion.
- Factors WhistleOut doesn’t account for, like device compatibility and customer service quality, matter to consumers.
While WhistleOut’s plan finder has a feature for checking coverage, WhistleOut appears to treat coverage as a binary thing—either you have coverage or you don’t. In reality, coverage quality is much richer. You can have mediocre coverage or strong coverage. You can have good coverage at your house but problematic coverage where you work.
While building CoverageCritic’s plan finder, I tried to account for things like prices, resource allotments, and coverage quality but kept in mind that these aspects of wireless service are complicated and often difficult to fully capture in simple models. While I can’t claim my tool is exclusively driven by hard data, I think my approach makes the tool more useful than competitor’s tools.
CoverageCritic’s tool makes predictions about coverage quality after drawing on geographic information provided by users. At the moment, state-level estimates of coverage quality are combined with user-provided information about population density. Population density proxies for coverage quality and is used to adjust state-level coverage estimates to arrive at location-specific predictions of coverage quality. In the future, I hope to refine the predictions of coverage quality by drawing on much larger data sets from carriers and network evaluators.
At the moment, the tool considers services from about ten carriers, and I plan to add more soon. The tool isn’t perfect, but it should be able to provide most consumers with a good starting point as they search for wireless providers.