# Data-driven SaaS for video surveillance [Carter Maslan](https://medium.com/@CarterMaslan?source=post_page---byline--35840f5a2146) 3 min read · Nov 16, 2017 Unlike fixed-price, on-premise, peak-capacity deployments, usage-based SaaS plans automatically reduce costs with dynamically provisioned compute capacity that’s shared across all cameras — even when they’re in different geographic locations. Software-as-a-Service (SaaS) has well-known advantages in lower upfront costs, painless continual upgrades, seamless integration with other systems, and increased end user usage (since access is simple from any computer, phone, or networked device). But video surveillance systems haven’t been able to exploit these SaaS advantages, because the video streams themselves are so massive that they can’t move through limited-bandwidth network pipes to connect to SaaS services in the cloud. ## Get Carter Maslan’s stories in your inbox Join Medium for free to get updates from this writer. When Security Operations Centers (SOCs) can instead use [Camio](/content/home/index.html) to _measure_ video surveillance events — with automatic options to control the movement of video and [metadata](/content/metadata/index.html) over the network — then the advantages of SaaS extend to video surveillance too. The [**Camio Data API**](https://api.camio.com/#data) makes it simple to work with the data generated by surveillance cameras to reduce costs while improving service. A single =ImportData() formula fetches comprehensive surveillance video data in real-time. Copy [this template](http://go.camio.com/sheet) or open [this redacted example](http://go.camio.com/sheetredacted). Camio not only replaces large upfront capital costs with lower monthly operational expenses but also enables subscription plans that optimize the allocation of budgeted funds. So even in scenarios where video surveillance is categorized as a cost (e.g. regulatory compliance) rather than an investment (e.g. police force multiplier), the allocation of budget is driven by data to achieve the lowest cost or highest return on investment. For example, a city with 100 cameras may have 14 for parks, 21 for building entrances, 25 for street corners, and 40 for busy intersections. Rather than purchasing equipment for peak-capacity video analytics for 100 cameras, the data on the number of important events observed by each camera — and the way in which each camera is searched — drives the assignment of optimal subscription plans for each camera. An outdoor park camera can be on a plan that’s 15x cheaper than a critical street intersection camera, because the park camera observes far fewer important events each day. A building entrance camera can be on a plan that’s 5x cheaper than a street corner camera for advanced search, because the real-time Deep Learning labeling of events is limited to people approaching or departing the particular entrance. And even the busy intersection cameras can be 100x cheaper to search for make and model of vehicle or license plate text when queries like “white truck san pablo ave 2pm to 2:30pm” trigger only on-demand analyses that are constrained to a short time periods of interest. The [Camio Dashboard](/content/dashboard/index.html) shows activity patterns across all the cameras on an account. Its primary purpose is to identify opportunities for [bandwidth savings](https://help.camio.com/hc/en-us/articles/115004413543-How-do-I-optimize-bandwidth-usage-) by adjusting the position, training, or zones of high-activity cameras. But all the data is also accessible via the [Camio API](https://api.camio.com/#data). There’s no programming required either— you can also import data directly into a Google spreadsheet with [this template](http://go.camio.com/sheet) (see a [redacted example](http://go.camio.com/sheetredacted)). Data-driven SaaS [plans](/content/price/index.html) — matched with dynamically provisioned service levels — finally extend to video surveillance all the cost, capability, integration, and usage benefits that other categories of software have enjoyed for years. Free the data! SaaS Video Surveillance Big Data Machine Learning