Sign in Subscribe

Tim Gaunt

Building a Personalisation Roadmap on a Realistic Budget

Most personalisation projects don't fail because of budget. They fail because nobody sequenced the work properly. Here's a practical twelve-month plan, split into three phases, that builds real personalisation capability without a platform purchase, a new hire, or a six-figure investment.

Running Products Alongside Client Work: Why It Took Us 15 Years to Get This Right

After 15 years building products at TSD, we're only now structuring this properly. Running an agency and running a product need different mindsets. Project completion vs continuous iteration. Clear deliverables vs never being "done". Here's what we learned the hard way.

A Cautionary Tale on Platform Dependency: Why Our Shopify App Went From Daily Installs to Zero Overnight

Platform products give you faster time to market and built-in distribution. They also give you existential risk. When Reddit changed API pricing, Apollo went from multi-million pound business to dead overnight. Here's what happened when our Shopify app installs stopped, and how we manage risk now.

From Napkin Sketch to App: How We Validate Product Ideas

How do you decide what products to build? We assume everything takes three times longer than estimated and markets are three times smaller than we think. Sounds pessimistic. Turns out to be fairly accurate. Here's the validation framework we use at TSD to avoid wasting months on ideas nobody wants.

The Personalisation Metrics That Actually Matter (And the Vanity Ones to Ignore)

Most personalisation reports I see are full of the wrong numbers. Opens, clicks, coverage percentages. They look like progress, but don't tell you whether revenue moved. Here's what to measure instead, and a simple framework to report it without needing a data science team.

The Cold Start Problem: Why Most Personalisation Fails Before It Starts

The most common reason personalisation underperforms isn't the tool or the team. It's that the data collection window closed before anyone realised it had opened.

We Built an Open-Source MCP Server for Productive.io (and You Can Have It)

We built an open-source MCP server that connects your AI assistant to Productive.io. Over 40 tools for tasks, budgets, and more. Free, MIT-licensed, and ready to use.

What Actually Is Web Personalisation and Why It Could Be Worth More Than Magic Beans

Personalisation isn't just saying "Hello Jack." It's a spectrum from basic recognition to invisible anticipation. Here's what each level looks like and the data you need to get there.

How "No Results" can create a revenue stream

Every search returning no results is a customer telling you what to stock. For one client, tracking zero-result searches created entirely new product categories. The best part? You don't need sophisticated analytics. Just a spreadsheet and a monthly review habit.

Your Sectors Plan Differently. Your Emails Should Too.

Your customers don't all plan at the same time, so why email them like they do? We tracked search behaviour across 11,000 retail accounts and found sector timing varies by months. Aligning campaigns to planning cycles lifted conversions 31%.