Category: data
-
DMA-CPA: a proposed margin-based acquisition budget heuristic
The DMA-CPA approach is a proposed budgeting heuristic: estimate the contribution available from a customer, reserve part of it, and use the remainder as an acquisition-spend ceiling. I find that more useful than chasing the lowest possible CPA without asking what a customer is worth. It is not a validated forecasting model or a guarantee…
-
Brand Impact Index: a proposed brand measurement framework
The Brand Impact Index (BII) framework presented here organises brand measurement into three connected areas: brand tracking, customer experience and business impact. Its purpose is to help a team decide what to measure and how findings should inform action. It is a proposed framework, not a validated industry index or a demonstrated predictor of growth.…
-
Using GPT-3-era tools at work: useful tasks and verification limits
This article concerns the GPT-3 and early ChatGPT period. It is a historical perspective, not a guide to the features or model lineup available today. My interest remains practical: where can a language tool save useful effort, and what must we check before acting on its output? GPT-3 and ChatGPT are different kinds of names…
-
What marketing mix modelling can and cannot measure
Marketing mix modelling (MMM) estimates how marketing activity and other factors contribute to an aggregate business outcome, such as sales. It can help answer questions about channel contribution, diminishing returns and budget allocation. It does not recreate individual customer journeys or automatically tell an advertiser which person, interest group or creative asset to target. Substantively…
-
Planning a marketing mix modelling project: data, checks and decisions
A useful marketing mix modelling project starts with a decision, not a promise to recover every conversion lost to tracking restrictions. MMM uses aggregate data to estimate how marketing and other factors relate to a business outcome. The practical work is deciding what the data can support, checking the assumptions and turning uncertain estimates into…
-
Four checks for effective data visualisation
A useful data visualisation helps someone answer a question. Looking polished is helpful, but it cannot rescue a chart that uses the wrong data or makes a comparison hard to understand. I find it useful to check four things: accuracy, clarity, design and practical use. They overlap, but asking about them separately can reveal problems…
-
An early AI-writing experiment—and why fluent text needs checking
This was an early experiment in AI-generated writing. The original article asked readers to spot the generated text, then disclosed its origin at the end. I now put that context first: the sample below was produced by an NLP tool, and the exact generator is not recorded here. It is preserved as an example to…
-
12 marketing automation predictions from 2021
These are twelve things I wanted to see more of in marketing automation when I wrote this in 2021. They are personal predictions and wishes, rather than a report on what subsequently happened or a list of capabilities every business needs today. Yes, I wanted machine learning and AI to take on more of the…
-
Data quality checks for marketing automation
Data quality can determine whether a marketing automation workflow does what you intended. Clean-looking records are not enough: the values need to be correct, relevant to the task and mapped to the right fields. Consider a lead form. If a phone number arrives in the first-name field, a personalised message can immediately reveal the mistake.…