We live for our KPIs. They’re the lifeblood of product growth. But, when they drop, executives start to ask what happened. You’ll often stop everything to focus on uncovering why, amidst a cacophony of possibilities - experiments, discounts, feature launches, and external influences - which is a painful process.
To decipher “why” your KPI dropped, first, recognize that there could be many possible reasons and that multiple factors can each play a part, complicating root cause analysis.
KPI drops typically consume a lot of your time. Understanding KPI drops manually or with traditional tools can feel impossible given the typical pace of business, depth of data, complex distribution of enterprise data, and that so many things are usually happening at the same time: features, new releases, marketing campaigns, bug fixes, etc. Relying on traditional correlation-based dashboards can leave analytics teams fruitlessly chasing guess after guess.
Imagine, an e-commerce platform sees a decline in Conversion during a New Year promotion. Perhaps, a surge of low-quality traffic driven by higher discounts overwhelmed the customer support team, increasing cart abandonment. How do you proceed?
A Framework:
To go from “guessing” to gaining “actionable clarity,” it helps to build a systematic framework to review and control for multiple factors:
- Seasonality: Make sure to account for the impact of holidays, weather, season, etc.
- Segments: Sometimes, there are hidden segments (clusters), such as campaign-driven users interested in a particular item.
- Marketing Campaigns: Are other concurrent campaigns having an impact?
- Other KPIs: Look for other KPIs that causally or mathematically affect your KPI. Building a KPI Tree is a good practice.
- Data Pipeline: Are there data process or quality issues, such as broken processes/ connections, stale data, schema issues, etc.?
- Product Launches: Are these affecting your KPI?
- Experiments: Are they misconfigured or slowing performance?
- Consider other factors: They may not be captured within your data.
Consider AI to free your analytics teams from the repetitive tasks involved in understanding and calculating attribution.
At Loops, our AI-driven platform automates the methodical and complete analysis of enterprise data sets, even across disparate data silos. It helps teams detect anomalies, build a KPI Tree, and identify root causes - find and rank top contributors by the size of their impact. Loops also push automatic alerts and summaries via Slack/ MS Teams or email, empowering teams to take swift, productive action.
KPI drops are inevitable. What sets successful teams apart is their ability to analyze and act effectively. By applying a systematic framework and leveraging advanced AI-driven analytics tools, instead of chasing KPI drops, your teams’ time can be focused on deep research and insights, and finding new growth opportunities.
