A Week in AI Marketing Analytics, Week 45
Here at Improvado, we’re changing the way marketers interact with data to uncover and explore performance insights. By introducing AI to their day-to-day operations via our AI Assistant, we constantly shift the focus from how to uncover insights to what insights they specifically want to uncover. AI Assistant is now in Beta, and we’re getting plenty of business questions to answer each day. The ‘week in AI marketing analytics’ series shares some stats to keep you on track with what AI can help you achieve.
Last Week’s Questions to AI
Improvado AI Assistant is an easy-to-use tool to get answer to any analytics question.
Since exact chats with our assistant are under an NDA, we went with categorizing the questions into use-case buckets. Here are the top five use cases from the last week:
- Campaign performance. Marketers and other business users asked AI for specific performance indicators. Something along the lines of ‘show me the daily combined ad spend from Google and Bing over the past week’ or ‘show me the top 5 campaigns that generated the highest ROI in the current quarter.’
- Aggregated campaign stats. It was important for marketers to rapidly understand how their campaigns were doing without digging into dashboards or reports. A good example would be: ‘show total ad spend broken down by ad platform.’ Also, ‘show how we are pacing with ad spend across our target geographies.’
- Comparative analysis. Often, it’s crucial to know how current performance compares with previous periods. ‘Compare conversion rates on Google Ads between October and September 2023’ would be a great example here.
- Averaging metrics across time periods. For monthly reporting and meeting preps, marketers asked the assistant to ‘show average CPC for January 2023,’ ‘show average ad spend across all platforms in 2023, broken down by month,’ and such.
- Cost per acquisition (CPA) analysis. To support their business decisions, marketers wanted to know which campaigns provided the best CPAs: ‘which campaigns in Google and Bing came with the highest CPA over the last 90 days.’
As data geeks, we see that the last week’s questions involved using multiple data sources and data tables to provide answers. This means that we helped marketers save some time and allowed non-technical business users to access precise performance data.
Stay tuned for more AI use cases in marketing analytics.
How it Works
Powered by a custom large language model (LLM) similar to ChatGPT, AI Assistant takes questions in plain English, translates them into SQL, and queries your dataset to provide you with instant insights. Your answer is then a new data frame, which you can further interact with by continuing your conversation with the assistant. Whether it’s asking for more granular insights or campaign advice, AI Assistant has got you covered.
In AI analytics, the quality of answers largely hinges on the quality of the underlying dataset. Improvado aggregates data from all your marketing and sales platforms, normalizes and performs quality assurance on it to prepare for further analysis. AI Assistant then acts as a data front-end that simplifies access and democratizes insights.