Tom Mansell
speaker
190 appearances
5 recordings
1 series
first heard Sep 2025
last heard 26 Sep
Tom Mansell’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
That I can give here is, you know, very simply, we were testing the impact of copy blocks on uh product lister pages, PLPs.
And the way that we measured the impact of that was to isolate, you know, control versus variant test groups, right?
We'd pick out some control pages that we leave as is, variant pages where we'd inject the copy blocks, and then you
You know, simply we just looked at the percentage share of revenue that came from organic search on those pages versus versus the control.
And it jumped from around 19% to 21% share of revenue, which was a 10% lift in revenue overall.
So that's an example, right, of where you've got a test and then you think about the measurement framework behind it.
And I think if you've got those things in place, it's an easier conversation to broach with the
business that isn't necessarily as set up for experimentation as it as it should be.
Yeah, I think it circles back to what we've been talking about, you know, understanding the risk, right?
And and my my advice would be start with start with the low risk areas, which is, you know, the reporting, the insight, the research.
And as you mentioned previously, if you can get those areas right, you can start to unlock some more of your time, which is, you know, super, super
Valuable and those areas are low risk from a performance perspective.
So if you can demonstrate efficiency drivers in the low-risk areas, then that sets you off on a on a good platform to roll that out further and unlock more experimentation budget within within a business.
I think where I talk about, you know, research, insights, engineering.
You know, one example of this that we're looking at a crowd is um a tool called uh brand AI.
So the way that works is querying the large language models through a range of different prompt patterns.
We're injecting the brand, the market, the category that they operate in, the different purchase considerations that um a customer would go through.
And we're extracting from the large language models, you know, what
How do you understand the brand?
How do you perceive the brand versus competitors in these different areas?
Showing 161–180 of 190 · page 9 of 10
← Previous
Next →