Examples of how to prompt for Benchmark tracking and analysis, segmented by metric tracking and analysis.
Last updated
All examples use Chat Explore, available in the Explore tab. Prompts utilize a dummy marketing dataset with columns for CTR, CPC, CPM, CTR_Benchmark, CPC_Benchmark, and CPM_Benchmark. Change prompt text as applicable to your data.
Tracking Against Benchmark
Identify campaign performance via Benchmark data.
Example Prompt:
Create a table of the top 25 performing campaigns based on CTR, CPC, and CPM vs their benchmark values.
Akkio Output:
Example Prompt:
Create a graph of CTR for Instagram over time, grouped by daily average, in 2024. Include CTR_Benchmark.
Akkio Output:
Tips:
Utilize live data connectors to have an automated dashboard that can track current metrics vs historic benchmarks.
Create benchmark values from industry-wide data to establish a competitive benchmark.
Create historical benchmark values in Prepare by creating new columns for the average value of each metric.
Example in Column Adjustment
Keep in mind, or reverse graphs, where lower values are preferred. Such as lower CPC and higher CTR.
Refer to Akkio's AI-Interpretation to check for ascending vs descending ranking by metric.
Benchmark Analysis
Identify and visualize relationships between metrics with a baseline comparison.
Exmple Prompt:
Create a graph of each Ad Type's CPM over time, grouped by day. Include average CPM_Benchmark for each date point.
Akkio Output:
Example Prompt:
Create a non-stacked bar chart of average CPC, CTR, and CPM for 2024 by Campaign Type. Include a red dashed line for the average Benchmark of CPC, CTR, and CPM.
Akkio Output:
Example Prompt:
What is the average variance of CTR, CPC, and CPM from their respective benchmarks?
Akkio Output:
Tips:
General work-flow - Prompt a graph as you would as usual to analyze your metrics. Include a sentence to add a dotted line for the metric's benchmark.
Keep consideration of variance by metric from the benchmark.
Example on how to prompt for variance listed above.
