Operating in today’s digital economy often involves dealing with an extensive network of third-party providers and partners.
Common types of partner networks include affiliates, vendors, suppliers, marketing platforms, and payment gateway providers. Partner networks involve tracking and analyzing data from multiple providers, each of which creates thousands of metrics and billions of events each day.
Affiliate Monitoring: Catching an At-Risk Account Just in Time
Companies that manage large networks with 1000s of affiliates often have a dedicated partner manager to handle the VIP accounts. Notifying these partner managers about issues with high-value affiliates is an incredibly important, albeit challenging problem to solve. Many companies have tried setting static thresholds to alert these managers about potential issues with affiliates – such as technical bugs, price glitches, or an account that may simply be at-risk or simply experienced an unexpected drop in traffic.
With such a large amount of data available to data-driven organizations, however, static thresholds often miss important changes. In order to deal with the inherent complexity of partner monitoring and to manage high-value affiliates, many of Anodot’s customers have turned to automated business monitoring.
Below are two examples of anomaly detection for partner monitoring in the adtech industry.
AdTech Monitoring: Sudden Spike in Video Ad Errors
For an adtech company, there was a sudden spike in video ad errors caused by a broken API, resulting in video ads not being displayed to viewers on a publishers’ site. Anodot alerted the adtech customer and the issue was forwarded to the R&D team. Due to the speed of the alert, the team was able to fix the API connection in minutes, rather than hours, saving their advertisers thousands in revenue.
AdTech Monitoring: Detecting Anomalous Drops In Traffic
Another example of partner monitoring is detecting anomalous drops in impressions for ads. The alert below was sent to an adtech company when one of their publishers was reporting fewer-than-normal ad impressions on their site. This was caused by an evening code update on the publisher’s website, which caused the ad to not display properly. The adtech company acted on the alert by proactively reaching out to the publisher, who was able to run a quick fix to resolve the issue.
Compared to the previous example, this anomaly was much more subtle and was likely caused by a small change on the partner’s side, for example a change in the location of the ad being served. This drop in partner traffic almost certainly would not have been caught with a static threshold, although due to the machine learning algorithm’s ability to automatically adapt to changing data, the anomaly was detected and the partner manager was notified in real-time. Again, this real-time detection and root cause analysis helped the company resolve an incident before it negatively impacted revenue.
Anodot’s automated solution can detect and group related incidents across the partner ecosystem at a speed, scale, and accuracy unattainable to manual monitoring and static dashboards.
Teams and partner managers can use these real-time alerts and correlations to uncover the underlying cause of the issue, ultimately leading to the fastest possible time-to-resolution.
If you’re interested in seeing what Autonomous Business Monitoring can do for your team, reach out to talk to us.