Insights
Perspectives from our work.
Practical lessons drawn from our engagements, each with actions leaders can take independently.
Operations before software
Document the operation before selecting software
Owners often begin by asking for a system. The most valuable work begins one step earlier.
A bus operator asked us for software to replace the notebooks and registers holding daily collections, fuel, wages, repairs and compliance documents. Before selecting anything, we documented how a typical day actually runs, from the first trip to the evening reconciliation.
That record shaped the system more than any feature list could have. Insurance, fitness and registration documents were easily missed, so the system now alerts the owner ten days before each one expires. The owner treats a poor day as something the next day must recover, so the daily profit figure carries losses forward rather than concealing them.
Software that reflects how the work is really done asks less change of staff, and that is what drives adoption.
Source: the fleet operations engagement described under Case studies, delivered and handed over to the owner.
- Document one typical day from start to finish.
- Record what the owner reviews each evening.
- List every document or payment that can lapse, with its date.
- Only then evaluate software against that list.
AI and reputation
What AI assistants read about your organisation
AI assistants read more than your visible pages. They also read the structured data behind them.
Most websites carry machine-readable data that visitors never see: the organisation's name, address, services and, at times, its awards. Search engines and AI assistants use it to describe the organisation.
In one AI-visibility audit, we found an award recorded in that data that could not be verified. We recommended removing it before an AI assistant repeated it as fact, since an inaccurate claim repeated by AI is far harder to correct than one on a web page.
AI visibility is not only a matter of being mentioned more often. It is a matter of being described accurately.
Source: the AI-visibility assessment described under Case studies, scored across six measures with a 30-day plan.
- Ask an AI assistant what it knows about your organisation, and note any inaccuracies.
- Ask your website provider to show you what its structured data states about you.
- Remove any claim you could not substantiate if a customer asked.
- Review again after every significant website change.
Advertising controls
Advertising budgets need controls the owner holds
Online advertising can exceed its planned budget without anyone noticing. A single limit is not sufficient.
Two organisations we work with advertise on Facebook and Instagram. We put three safeguards in place for them: an overall limit that we monitor independently, so it holds even if a platform setting is wrong; a daily threshold that pauses campaigns and asks the owner before further spend; and an approval that only a person can give.
The last of these matters more than it appears. Some email systems open links automatically to check them for threats, so we ensured that opening the email never approves anything. The owner must make the decision deliberately.
The principle applies to any organisation: spending should stop and ask a person, rather than continue until someone notices.
Source: the advertising safeguards described under Case studies, live for both organisations.
- Set and record the maximum you will spend on any campaign.
- Check actual spend against it weekly, not only the platform's own total.
- Require a person's approval before spend exceeds the daily limit.
- Ask your advertising provider what happens if a limit is set incorrectly.
Reading your traffic
Confirm that your visitors are real before you act on the numbers
A healthy-looking visitor count can hide a different story.
For one business, website traffic appeared healthy while few of the visitors appeared to be genuine. We compared engagement with session duration to separate automated traffic from real visits.
Once the automated traffic was set aside, a real overseas audience became visible, and so did revenue tracking that was active but had never been configured correctly. The client's effort could then move to the channels that matter.
A visitor count measures activity. It does not measure interest, and decisions about where to spend should rest on interest.
Source: the traffic analysis described under Case studies, with findings delivered to the client.
- Compare the engagement rate with the session duration for each channel.
- Separate automated visits from genuine ones before reporting a total.
- Check where the genuine visitors are and what they do.
- Confirm that revenue or enquiry tracking is configured correctly, not only switched on.
Moving a website
Move a compromised website without interrupting the business
When a website has been tampered with, the instinct is to switch it off. Careful sequencing avoids that.
A food producer's WordPress website had been compromised, with malicious code hidden in old pages. We worked on a copy of the site, removed the injected code, and recovered pages that had been lost from web archives.
We then rebuilt the site and planned the move of the domain and hosting in an order that avoided any downtime, verifying the new hosting before the domain was switched.
A compromised website is a reason to plan the move carefully, not to rush it.
Source: the website recovery described under Case studies, live in production on its own domain.
- Work on a copy, and leave the original in place until the new site has been checked.
- Search old pages, not only recent ones, for code nobody on your team wrote.
- Look in web archives for lost pages before assuming they are gone.
- Switch the domain last, once the new hosting has been tested.
Reaching buyers
When advertising cannot find your buyers, reach them directly
Advertising platforms target by interest and category. Some buyers fit neither.
A food business wanted to sell to institutions as gifts, a buyer that advertising categories do not describe. We built prospect lists by region and category and wrote a separate outreach message for each category.
We then moved the budget from paid advertising towards direct outreach in phases, not all at once.
The lesson is to begin with who the buyer is and to choose the channel afterwards.
Source: the outreach campaign described under Case studies.
- List the organisations that could buy, by region and category.
- Write one message for each category, not one for everyone.
- Move budget in phases, and review the results before each move.
- Keep paid advertising for audiences the platform can describe.
To discuss how these lessons apply to your organisation, .
Contact
Discuss your organisation's priorities with our partners.
We respond to every enquiry, and we will tell you plainly whether we are the right firm to help.