As Google rolls out AI-assisted performance reporting, iCrossing UK’s Edouard Payne road-tests the current offering.
Google's AI performance reporting: How does it work, and is it any good? / Steve Johnson via Unsplash
Earlier this year, we were expecting Google to incorporate AI performance reporting in October. The integration, we said at the time, would help marketers improve the time it takes to generate insight from marketing performance by allowing you to conversationally ask questions about the marketing data you collect.
Since then, Google has slowly been releasing this part of the feature across their partners. So: How’s it looking so far?
A decline in data-driven business?
User journeys are becoming increasingly complex; Google itself has said that users can experience 20-500 touchpoints before a conversion. Amid such complexity, it can be hard to decide how much should be invested in each of these touchpoints. In 2021, one study found only 24% of respondents saying that they thought their organization was ‘data-driven’ in the past year, a decline from 37.8% the year before.
To combat this, companies have started to combine data from different touchpoints into a single source of truth – in a data warehouse or ‘lakehouse’. This simplifies the time-to-insight; instead of trying to find your data in multiple platforms you now have a singular platform.
The key challenge for marketers in using the platform is that it requires new skills to extract and segment, such as SQL (structured query language, a data-management programming language). These skills are not traditionally a prerequisite for marketers, so data analysts can be a useful source of additional insight – if you have them at your disposal.
Google’s AI solution
Google’s data warehouse BigQuery and new AI functionality can address these challenges by giving you the tools to interrogate the data that you collect. The process should be simple: first, you ask a question about your marketing data, just as you would ask a data analyst. Then, the AI generates the SQL and other outputs, based on how the data is stored for your business. It may also (if needed) produce a graph as a visualisation.
For example, we asked the AI a typical marketing question: “Give me spend broken out by platform in 2023”. From this information, it’s able to understand the semantic meaning of the questions and what columns to use from the database. It then creates SQL to extract the data.
With this data we can then ask it to create visuals – the tool is flexible enough to change the colors of the chart and rename the columns based on our questions. In our test we asked it to color the chart in ‘barbie’ pink.
It’s not limited to just bar charts; in another example we combined historical GA3 & GA4 data, asking for a line chart. It understood the need to create a multi-axis chart to easily show year-on-year growth.
This shows, so far, great flexibility with the data that you collect and visualize, cutting time to deliver insight by 80%-90%. The AI can also explain the SQL code that it generates.
Other potential use cases
Remember, the feature is still in preview mode. Expect further improvements in its ability to understand, summarize, and plot data. Still, it’s not a silver bullet in answering your marketing needs – but it does serve as a useful tool in improving your data-driven capability.
Suggested newsletters for you
Daily Briefing
Daily
Catch up on the most important stories of the day, curated by our editorial team.
Ads of the Week
Wednesday
See the best ads of the last week - all in one place.
The Drum Insider
Once a month
Learn how to pitch to our editors and get published on The Drum.
It’s increasingly important for marketers to consolidate their data into a single platform, to benefit from further advances in generative AI. Invest now in such a solution to futureproof yourself and stay competitive.
At iCrossing, for example, we use sentiment analysis to delve into the emotional depth and sentiment expressed in comments on social media advertisements and posts. This approach allows us to gauge public reaction and adjust our strategies accordingly. Similarly, we apply sentiment analysis to evaluate and report on the overall perception of a brand in search engine results pages (SERPs), ensuring we maintain a positive online presence.
To streamline our data analysis, we employ AI categorization techniques to systematically organize keywords and web pages into coherent groups based on categories, themes, intent, and stages of the user journey. This method enables us to generate insightful reports that highlight the performance of these segments, leading to more informed and strategic decisions.
Finally, through generative AI, we can craft timely and relevant reports that are automatically distributed via Slack, keeping our team and clients updated on the latest industry developments.