“In just an hour, I built a UI to interrogate my data, and it handled most of the heavy lifting for a client project."

Chris Marshall, Director & Co-founder, OnState.

Optimising Ecommerce Data with AI: Real-World Applications

Yes, we talk about AI a lot on the podcast. It's inevitable, AI is weaving its way into so many ecommerce processes and tasks.

This episode is highly practical.

We cover real-world examples of how AI tools are being used to speed-up product data tasks whilst reducing the need to rely on expensive licences for specialist tools.

Summary

Ecommerce businesses are increasingly turning to AI to enhance their data management processes. The pod explores how AI tools are being used to clean, enrich, and structure product data, providing real-world examples that highlight their practical applications.

The Build vs. Buy Dilemma Is Being Reframed

Businesses often face the decision of whether to build custom solutions or purchase existing platforms.

In the context of AI for product data, building allows for tailored solutions using tools like Google Sheets and AI models such as ChatGPT for tasks including data transformation and HTML cleaning.

On the other hand, buying involves using specialized AI-enabled tools or outsourcing, which can save time but may incur higher costs.

Practical AI Strategies Discussed:

1. DIY data cleaning: AI models can automate data cleaning tasks, such as reformatting unstructured HTML and standardising attributes, saving significant manual effort.
2. Automating data structure: AI can analyse complex datasets, infer attribute types, and suggest categorisation rules, streamlining the setup of dynamic product groups.
3. Hybrid approaches: combining DIY methods with outsourcing can optimise resources, allowing businesses to handle unique projects efficiently.

Tune in to hear how AI is transforming data migration and management by automating previously manual tasks, increasing speed and allowing for continuous learning.

Chapters

[00:30] The Build vs. Buy Debate in AI Data Management

[03:20] AI in Data Migration: Practical Use Cases

[06:15] Transforming Data with AI Tools

[09:20] The Role of AI in Content Management

[12:20] Engaging with Data Structures

[15:00] Building Custom AI Tools for Specific Needs

[17:45] Tactical Middleware: A New Approach

[20:35] Speeding Up Data Transformation Processes

[23:20] Validating AI Outputs and Managing Expectations

[26:15] The Future of AI in Ecommerce Data Management

"It's not just about selling more; it's about selling smarter and more sustainably."

Oscar Rundqvist, CEO, eComID.

Summary:

High return rates are a significant challenge in fashion ecommerce, leading to financial losses and environmental harm. This podcast explores how innovative technology like eComID is helping the industry by reducing returns, personalising customer experiences and promoting sustainability.

The problem...

Returns are costly; retailers face expenses from shipping, inspecting and repackaging returned items. Additionally, unsellable returns often end up in landfills, contributing to environmental damage. It's estimated that returns cost retailers between $21 to $46 per item, highlighting the need for effective strategies to minimise them.

Change is needed

Retailers need better metrics to understand customer behaviour.

Traditional metrics like average order value (AOV) can be misleading. Instead, brands should focus on "kept order value," which reflects the revenue retained after returns. This shift in focus can lead to more accurate assessments of profitability and customer loyalty.

Emerging solutions like eCom ID offer innovative ways to tackle return issues. By creating a digital shopping passport, eComID stores individual preferences related to size, style and fit. This enables precise pre-purchase guidance and personalised size recommendations, reducing the likelihood of returns.

Data from post-purchase behaviour and return reason codes are invaluable for improving size accuracy and product fit. Brands like H&M have learned that systemic issues, such as inconsistent sizing, drive return rates. By incentivising detailed feedback, brands can refine their offerings and policies.

Innovative brands are experimenting with personalised return fees based on shopper history. This approach encourages mindful purchasing and supports sustainability by reducing unnecessary returns.

Reducing returns is not just about cutting costs; it's a step toward a more sustainable fashion industry. By utilising personalised, data-driven tools, brands can enhance customer experiences and build lasting loyalty.

Chapters:

[00:30] Introduction to Tech and Sustainability in Fashion

[02:40] Understanding eCom ID and Its Solutions

[07:10] Insights from H&M: The Returns Challenge

[12:45] Data Accuracy in Returns Management

[15:35] The Role of Size and Fit in E-commerce

[20:20] Measuring Success: Average Kept Order Value

[24:15] The Environmental Impact of Returns

[29:50] Future of Ecommerce: Agentic Commerce and Virtual Try-Ons

Why did we think this was an interesting episode?

Paul & James are regularly involved in ecommerce redesign projects, either in an advisory capacity or helping drive the design thinking.

This episode explores the reasons why brands decide to invest in a redesign:

• Brand refresh or a full rebrand.
• Brand elevation of the online UX e.g. premium positioning.
• Improved user journeys to fix legacy constraints.
• Outcome focused e.g. fix navigation and browse journeys.

It then teases out the justifications for redesign projects, sharing views on how design can and should be measured objectively.

James & Paul also dissect the intangible goal of many design projects: to elevate the brand positioning, to create a premium look & feel.

Goals like this need clear definition and framing to ensure the design outputs work towards a clear vision and execution. They also need tangible measurements of success, even if they're not conversion focused.

The key take-away is that design has to be measured, and the metrics you use need to be agreed upfront. If there are no hard & fast commercial success metrics like conversion and AOV, then take a sensible approach to measuring customer impact, for example customer satisfaction & NPS.

Chapters:

[00:30] Introduction to Redesign Metrics

[03:40] Understanding Brand Elevation

[06:10] Balancing Design and Ecommerce

[09:00] Defining a Premium Experience

[12:30] Measuring Redesign Success

[15:25] The Role of User Testing

[18:15] Navigating Redesign Challenges

[21:10] The Importance of KPIs

[22:46] Final Thoughts on Redesigns