How NextRoll’s Technology Engages Customers Effectively
NextRoll's Tech Stack Series #4
Businesses and marketing teams of all sizes have the same goal of effectively reaching their audience through advertising and outreach. What defines the winners in this battle for that attention online is having the right marketing technology to deploy digital ads for the right price. There’s no one way to grab that attention, but there is one tool that can help businesses of all sizes spend their marketing dollars effectively and efficiently.
At NextRoll, we’ve developed a powerful Demand Side Platform (DSP) that allows our users to put the right ads for the right products in front of the right customers at the right time. Our tech stack also gives thousands of AdRoll and RollWorks customers the ability to track an ad's effectiveness – and most importantly, if it leads to a conversion or sale.
How does our technology differ from others in the industry? Well, like most marketing tech companies, NextRoll uses data to create anonymous profiles for web users. But we go a step further, as our Customer Data Platform (CDP+) incorporates data from first- and third-party data sources. For businesses that use our AdRoll platform, we can include account info or data from e-commerce platforms like Shopify. At the same time, businesses that use our RollWorks platform can combine data from a client's customer relationship management (CRM) and marketing analytics (MA) systems. And what really makes us stand out from our competitors is the billions of proprietary data points we use specifically to help growing small and mid-sized businesses identify, track, and place the ideal ad for their customers.
Better engagement through BidIQ
Ads are a vital part of the internet landscape today, and users see them everywhere. An ad could be on a news site, before a video plays, between rounds of a mobile game, or while viewing social media posts. It’s an integral part of online life, but most people are unaware of the flurry of activity happening behind the scenes while a web page with ads loads. It works like this: web page publishers partner with sites that auction off ad space. As pages load, the supply-side platform notifies interested bidders about the auction, and the winner gets to place the ad. More desirable websites (or prominent placement of ads) command a higher price.
This process happens incredibly quickly, and in those fractions of a second, our Machine Learning-powered system weighs several different factors in deciding if it should make a bid and for how much. These include, but are not limited to:
Targeting: Is the user meet the criteria of the AdRoll or RollWorks segmentation? What are the chances they will eventually reach a favorable outcome, like a sale or filling out a lead generation form on a company's webpage?
Bidding: We can reach 98% of the open web, so our BidIQ algorithm needs to evaluate everything from the quality of the site where the ad will be placed (is it a premium site like the New York Times or a less well-known blog), the location of the ad (i.e., top or bottom of an article, before a popular video, etc.), and time of day before making a bid. It assesses the user and their past online behavior to determine a fair price to bid for the spot if there is a reasonable chance it could lead to a conversion, which could be as low as a few cents.
Budget & Pacing Optimization: The system looks at what the customer's campaign entails and will work to allocate ads in the most optimal way possible. Say a customer has $500 to spend; it will decide how much should go to the Open Web or places like Facebook, LinkedIn, or TikTok. Pacing is always important. If the campaign is set to go for a month, the system works to ensure it doesn't spend all the money in the first few days.
Ad Selection: Many customers have multiple ads designed to target different segments. Our system evaluates the user to decide which creative element would work best to entice interaction and, eventually, a favorable outcome.
Our system excels at these calculations, and this focus helps our customers achieve results. Every time the system enters a bid, it accumulates data on the experience. It looks at how the user responded, how much it cost to win the bid, did it result in an eventual sale. It uses this information and learns how to effectively reach users at the individual level, not just the group. On average, it handles a staggering 120 billion auction requests daily, or nearly 1.4 million per second (that’s about 20X the average daily trading volume of the New York Stock Exchange)!
Evolving Our Tech to Meet Industry Changes
Our teams developed BidIQ in-house at NextRoll, and it’s constantly refined and optimized. New data points are constantly added, including feedback from successful and unsuccessful bids, so it’s continuously learning while staying ahead of changes in the industry.
In 2014, the online advertising industry started using header bidding, which allows publishers to offer ad space to multiple advertising exchanges simultaneously. This sent bidding costs higher because buyers could find themselves bidding on the same ad space through two or more exchanges, effectively competing against themselves. This is why we use our Supply Path Optimization (SPO) process to recognize duplicate auctions and eliminate any intermediaries that can contribute to a higher price. Today, about 90% of BidIQ purchases are made directly from the primary seller (source) of publisher inventory, which maximizes the spending for our AdRoll and RollWorks clients and protects them from overbidding.
For any growing business, ineffective marketing efforts waste time and money. Companies that can get their messages in front of the right people engage better with their ads, ultimately leading to higher sales. NextRoll is democratizing digital marketing to help companies with limited budgets, regardless of whether they are a B2B or D2C operation. Our technology allows our partners to leverage an immense amount of data and ensure their messaging hits the mark for the best price.