📉 Enterprise AI · 🙋‍♀️ User Data

IBM - b2b Saas product analytics

TL;DR | Created a whole new internal framework for capturing continuous user insights from B2B product data. Work led to v1 playbook for utilizing insights to then prioritize new UX requirements for higher success rate for customer acquisition.

Duration

Team

Tools

10 weeks

IBM Automatoon Platform (1 PM, 1 Engineering Architect)

Amplitude, Figma

The Project

IBM Software was venturing into product instrumentation as a B2B entity, starting with its in-house Automation Platform (new initiative). The goal is to get customer data live across the discover, try & buy experience for the team to continuously improve our customer acquisition funnel. I was responsible for setting up a data analysis framework that the team can use for its upcoming releases so that once the live customer data come in we are able to measure it against our KPIs and rapidly pinpoint which area in the UX needs prioritization.

My Role

Lead strategist for product marketing analysis framework

The Challenge

Because this was a novel initiative, there was almost no prior work done on product instrumentation (other than IBM’s acquired SaaS products that came with their own native telemetry). This meant I pretty much had to figure it out on my own.

My Approach

  1. Find and meet internal SME’s

    • Point to resources / read.me’s on existing telemetry efforts

    • Regular meetings for feedback on WIP

  2. Gather similar frameworks

    • Use ideation templates from different purposes (i.e. jobs-to-be-done)

    • Repurpose them into new framework for marketing data across product funnel phases

  3. Define variables

    • Horizontal > key user events by sales funnel phases

    • Vertical > phase, UX e2e, user events, etc.

  4. Set KPI’s and compare with benchmarks

We would be observing customer data from an real-time dashboard using Amplitude, like this one. (Source: amplitude.com)

Impact & Outcomes

Created a new framework for continuous user insights and product analytics for legacy B2B funnel.

IN HINDSIGHT

If I were to have taken on this initiative today, here is how I would have worked with my team using AI tools.

  1. Generating example frameworks

    • Could generate example frameworks to get inspiration to help things get started

    • Use incremental prompting to adjust based on our own constraints and needs


  2. Gathering benchmarks

    • Use agents to help research industry benchmark metrics at different funnel phases to fare how our product is doing

    • ***Understand risks of using AI to validate research...clearly audit the sources of where AI is getting its validations.


  3. Simulation models

    • Could build prediction models to test how well the current try-and-buy UX would fare

    • Could gather existing data from sales teams to train model