The advent of AI tools and LLMs is bringing with it one of the most exciting and disruptive changes to marketing analytics in decades - the ability to talk with your marketing data using natural language.

This article explains some of our reasoning behind deciding to integrate AI into our experiment analytics, and why we think natural language is the future of marketing analytics.

Why use natural language?

Imagine going into GA4 and generating a trend line showing the number of conversions from all blog pages in the last month. How many people at your company could do this right now in less than 20 seconds?

Marketing analytics tools, databases, event data and BI tools are notoriously difficult for the average marketer to understand, never mind other business stakeholders (product teams, customer success leadership etc). For complex tasks, UIs quickly become hard to use with rows of buttons and menus, input fields, filters, conditions and regular expressions.

The ability to use natural language and chat with your data removes all of the challenges associated with traditional graphical user interfaces - no platform certifications, no complex interfaces and no need for regular expressions or SQL.

Put simply, chat is a far better interface to complex data and reporting than, for example, the GA4 UI. It's a more natural way for people to interact with computers, and data.

Natural language in Growth Method

We recently introduced the ability to pull data from analytics tools in Growth Method using AI and natural language, removing a large part of the complexity in teams running marketing experiments within our platform.

Growth Method can now connect to common analytics tools (GA4, PostHog, MixPanel, Amplitude etc), write the necessary API query on the fly, request the data from your provider and add it to your marketing experiments automatically.

Who else is exploring AI for analytics?

A lot of companies! Below are a few examples.

1 Text-to-SQL

There are already various Text-to-SQL implementations that are streamlining data retrieval for non-technical users by turning everyday language into structured SQL queries:

We took the rise in availability of Large Language Models (LLMs) as an opportunity to explore whether we could assist our data users with this task by developing a Text-to-SQL feature which transforms these analytical questions directly into code.
Pinterest Engineering

Below is an example showing how scale.com can transform and standard natural language query into a complex SQL instruction in order to return the exact data requested by the user.

Image source

2 Cube

Cube acquired Delphi Labs, one of the first companies to use third party semantic layers for self-serve analytics platforms. Below is an example of the Delphi GA4 integration:

Image source

For more information see https://delphihq.substack.com/p/ai-for-google-analytics.

3 Prototypr.ai

Gareth Cull at Prototypr has built a conversational reporting interface to fetch data from multiple sources, including Google Analytics, Google Ads and Google Search Console. See the demonstration video below.

For more information see Gareth's announcement post on LinkedIn.

4 Microsoft

Among BI and analytics vendors, Microsoft arguably is ahead of the curve when it comes to AI, likely as a result of its partnership with OpenAI.

To-date Microsoft has added Copilot, its AI application assistant, to Microsoft Excel, Dynamics 365 Customer Insights and Power BI.

Image source

5 MixPanel Spark

MixPanel Spark allows MixPanels customers to answer their product, marketing, and revenue questions with natural language, effectively enabling companies to chat with their data.

At the time of writing Spark leverages OpenAI’s GPT-3.5 Turbo model, a smaller, more refined version of GPT-3.

For more information see the announcement blog post here https://mixpanel.com/blog/spark-bringing-generative-ai-to-mixpanel/.

6 Amplitude

Similarly to MixPanel above, Ask Amplitude enables you to go from business questions to insights in seconds - without any understanding of the Amplitude product.

"Powered by large language models, Ask Amplitude understands your product, your taxonomy, and your language to frame your analysis in the right way. Just type a question, and we'll share a chart with the answer you need."
https://amplitude.com/ai

Amplitude AI aims to make what complex data science problems, such as anomly detection and root cause analysis, accessible to everyone.

7 Looker Studio

Google recently introduced Conversionational Analytics as part of Gemini in Looker enabling users to ask questions about their data with everyday language.

"Conversational Analytics is a dedicated space in Looker for you to chat with and engage with your business data, as simply as you would ask a colleague a question on chat. Now, your entire company, including business and analyst teams, can chat with your data and obtain insights in seconds, fully enabling your data-driven decision-making culture."
https://cloud.google.com/blog/products/data-analytics/introducing-gemini-in-looker-at-next24

At the time of launch Conversational Analytics works with a limited number of data sources (including BigQuery and Google Sheets) and requires a Looker Studio Pro license.

Note: this starts at around 23 minutes for the conversational analytics demo

About Growth Method

In the future the majority of people will talk with their data. We believe over the next few years companies will double-down on API-first strategies in order to participate in data exchange, and so at Growth Method we're already building for this future.

Growth Method is a the only work management platform built for growth marketers.

Learn more at on our homepage, connect with me on LinkedIn or Twitter, or book a call here.