Notes on work, home, and trying to do both properly.
This is where I write about the things that have actually stretched me. Digital delivery, product data, analytics, security, AI, focus, IVF, becoming a dad, and learning how to stay useful when life is already full. The context is deliberately practical: regional businesses, real customers, family appointments, early starts, tired evenings and work that still needs to be done properly.
The aim is simple: honest notes with enough detail to be useful. No polished thought leadership. No pretending everything was obvious at the time.
I used to think balance meant finding a way to fit everything in.
Work, family, training, hobbies, friends, learning, house jobs, sleep, messages, life admin. All of it. Somehow. Every week.
I am not sure I believe that anymore.
Since becoming a dad, I have started to realise that balance is not about keeping every plate spinning at the same speed. It is about being honest about which plates cannot drop, and which ones can slow down for a bit.
Family is non-negotiable.
Work is non-negotiable.
That does not mean they always get the best version of me. I am human, tired and still working it out. But they are the things that need protecting. Being present at home matters. Providing matters. Doing my job properly matters. Being someone my partner can rely on matters.
Everything else has to flex.
The gym matters. It helps me switch off, clear my head and keep some structure when everything else feels a bit chaotic. Hobbies matter too. So do friends, messages, side projects and the bits of life that make me feel like me.
But they are not above family.
They are not above showing up at home.
And they are not above doing the work that helps keep life steady.
That has probably been the hardest part to accept. I like doing things properly. I like progress. I like feeling like I am moving forward. So when the gym becomes less consistent, or hobbies go quiet, or messages take longer to reply to, it can feel like slipping.
But maybe it is not slipping.
Maybe it is prioritising.
There is a difference between giving up on something and letting it breathe for a season.
The gym can wait a day. A hobby can wait a week. A message can wait until my head is actually in it. A side project can sit half-finished without meaning I have failed.
But my son is only this small once.
My partner should not feel like she is competing with the leftover version of me.
And work still deserves focus when I am there, not because it matters more than home, but because doing it properly helps protect the life around it.
So maybe the lesson is not how to do everything.
Maybe the lesson is learning what order things go in.
Right now, that order is clearer than it has ever been.
Family first.
Work done properly.
Everything else fitted in where it can, without pretending there is an endless amount of me to go around.
That is not perfect balance.
But it is probably a more honest one.
· Small business web
Small business websites should do a job, not just exist.
A lot of small business websites feel like they were made because someone said the business needed one. There is a logo, a few photos, a contact page and not much evidence that anyone has asked what the site is supposed to do. That is especially obvious for smaller operators, where the website often has to carry a lot of trust before anyone phones, books or visits.
For pubs, food businesses and independent operators, the basics matter more than people think. Are you open? What do you serve? Can I book? Where are you? Is the menu current? What makes you worth choosing over the place down the road?
The clever bit is not always technical. A clear one-page site with useful content, fast loading, good photography, proper local SEO, accessible markup and simple tracking can do more than a bigger site that hides the answers.
Good web work often feels obvious afterwards. Fewer dead ends. Better structure. Clear actions. Content that sounds like the business. Enough measurement to know what is working. Nothing fancy for the sake of it.
· Work and family
Paternity leave has a delivery problem.
Becoming a dad made me think differently about how work talks about family, especially when the real world is appointments, forms, rota juggling and trying to be useful at home as well as at work. There is a lot of warm language around bonding, support and being present, but the practical design often feels much thinner than the sentiment.
The early weeks are not a holiday with a baby in the room. They are recovery, feeding, admin, sleep maths, appointments, visitors, washing, worry, joy and a strange amount of sterilising. It is full-time work, just not the kind that fits neatly into a calendar invite.
For dads and partners, the gap is often between what people say matters and what the system makes possible. Being useful at home takes time. Learning the baby takes time. Supporting your partner properly takes time. You cannot compress all of that into a couple of neat weeks and call it sorted.
I do not think every answer needs to be complicated. More flexibility, better manager guidance, phased returns and a bit less awkwardness around men needing to be genuinely present would all help.
The business case is not fluffy either. People remember how work treated them when life got real.
· New dad
New dad mode: love, snacks and very little sleep.
Becoming a dad is the biggest context switch I have ever had. One day you are a fairly normal adult with routines, plans and a loose idea of what tired means. Then your baby arrives and suddenly the whole day is built around feeds, nappies, naps and tiny noises you are not sure whether to Google.
People tell you it will be tiring. They are right, but they cannot really hand you the feeling in advance. It is physical, emotional and weirdly practical. You become very interested in clean bottles, spare muslins and whether you can make a sandwich with one hand.
The best lesson so far is that useful beats perfect. Do the next job. Make the bottle. Take the photo. Let your partner sleep. Say sorry quickly when you are both knackered. Accept that some days the win is everyone being fed, clean enough and still kind to each other.
It has changed how I think about ambition too. I still want to do good work and build useful things. I just care a lot more about coming home with enough left for the people who get the real version of me.
· AI and workflow
LLMs are better as boring tools than magic tricks.
The more I use LLMs, the less interested I am in the impressive demos. The useful work is normally much plainer than that.
In a practical business setting, the value is not asking a chatbot something vague and hoping it sounds clever. It is helping people find the right product, improve messy content, summarise internal knowledge, draft a better first response, spot missing information and reduce repetitive work.
That only works when the basics are in place. If product data is weak, rules are unclear and nobody agrees what success looks like, adding AI just gives you a faster way to expose the same problems.
The good use cases are usually specific. Take this input. Use this context. Follow these rules. Show your source. Help the person get to a better answer quicker. That is less exciting than the big promises, but it is also much more likely to survive contact with real work.
My current view is simple: AI should make the workbench better. It should not become the person pretending to know how the whole workshop runs.
· IVF and life
What IVF taught me about uncertainty.
IVF taught me that waiting can be active, especially when everything runs on appointments, phone calls, letters and dates you cannot control. Waiting for appointments. Waiting for calls. Waiting for results. Waiting while trying not to read too much into every small thing. Waiting while wanting to be positive, but not wanting to tempt fate by saying anything too loudly.
I am wired to solve problems. Give me a broken process or an awkward system and I will start looking for the thread to pull. IVF does not give you that kind of control. A lot of the time, the best thing you can do is be steady and not make your own need to fix things the centre of the room.
That is harder than it sounds. Support can look very ordinary from the outside. Make the tea. Remember the details. Do the admin. Listen without jumping straight to solutions. Give your partner space to feel whatever the day has brought.
There is a work lesson in there too, although I did not see it at first. Some difficult situations do not need more force. They need patience, clarity and people who can stay calm without switching off emotionally.
· Data and ecommerce
Product data is not admin. It is customer experience.
Product data looks like admin until customers have to rely on it. Then it becomes search, filtering, confidence, substitutions, recommendations, compliance and support. In practice, it becomes a large part of the website.
The problem is rarely just a messy spreadsheet. It is usually unclear ownership. Who decides the naming standard? Who approves attributes? Who checks supplier changes? Who knows when a field is fine internally but not good enough for a customer journey?
Without those answers, teams end up fixing the same issues in different places. Search gets patched. Pages get rewritten. Reports need caveats. Customers take the long way round. None of that feels like a data problem in the moment, but it usually started there.
The practical answer is not glamorous. Clear definitions. Named owners. Approval routes. Change history. Supplier expectations. A shared view of what good looks like. It is slow foundation work, but it is the work that makes better ecommerce possible.
If the data is weak, the clever tools have less to stand on.
· Enterprise delivery
Enterprise change needs release control, not heroics.
One of the easiest mistakes in digital change is treating a major platform update like a visual refresh. The visible work gets all the attention because it is easier to react to. New pages. Cleaner components. Better journeys. Everyone can point at it and have an opinion.
The risky work is usually quieter. Data migration. Checkout behaviour. Account logic. Rollback planning. Content ownership. Testing routes that nobody notices until they fail. That is the bit that separates a confident release from a brave one.
The hardest part is keeping the scope honest. As soon as people can see progress, they can also see all the things they would like to add. Some of those ideas are good. Some are even right. That does not mean they belong in the release.
What helped me was being clear about the line. If it protected customers, transactions, data, support or rollback, it stayed close. If it was a preference, improvement or nice extra, it went into the next phase. Not because it did not matter, but because uncontrolled scope is how good work turns scruffy.
I have learned to trust release discipline more than big launch energy. A calm go live is rarely an accident.
· Data maturity
Data maturity is a culture problem before it is a tooling problem.
It is tempting to treat data maturity as a system choice. Buy the platform, create the fields, connect the integrations and wait for everything to become cleaner. I wish it worked like that.
The system matters, but it can only enforce the decisions the business has already made. Who owns product names? Which attributes are mandatory? What counts as good enough? Who approves supplier changes? What happens when internal data works for operations but falls short for customers?
Those are cultural questions as much as technical ones. They need ownership, shared standards and teams who understand that data quality is not somebody else's admin. It is how the business describes itself to customers, colleagues, suppliers and future systems.
The awkward bit is that data maturity asks people to slow down before they can speed up. Define the terms. Agree the process. Remove duplicate effort. Make the boring decisions visible.
Better tooling can make that work easier. It cannot make the business care. That part has to come first.
· DevOps and delivery
Good DevOps is mostly good manners.
Tools help, but they do not save you from unclear thinking. A board full of tickets can still be chaos if nobody knows what the work means, why it matters or when it is done.
The best DevOps habits are basic, but not always easy, particularly in smaller teams where everyone is close enough to assume the detail is obvious. Write the ticket so someone else can understand it. Separate the problem from the solution. Link the evidence. Add the test notes. Be clear about what is in scope. Do not hide decisions in private messages if the whole team needs them later.
That might sound boring, but boring is useful when delivery gets busy. It stops the same question being asked five times. It gives testers a fair chance. It helps managers understand risk without needing to sit inside every conversation. It makes handovers less painful.
I have become less impressed by busy boards and more impressed by tidy thinking. Good structure should reduce noise, not create a second job feeding the tool.
In the end, good DevOps is a courtesy to the next person who has to pick up the work. Sometimes that person is you, two weeks later, wondering what on earth past-you meant.
· Security and operations
Security improvements do not need to sound scary to matter.
Security work can be hard to communicate because the language gets dramatic very quickly. Threats, attacks, blocks, hardening, incidents. All true enough, but not always helpful when you are trying to build support for sensible improvements.
The better framing is often business protection, especially for organisations where the website is not just marketing, but part of how customers get served. Keep customers online. Reduce avoidable risk. Make changes traceable. Stop obvious rubbish reaching the application. Know what you would do if a release went wrong.
That is where the practical pieces matter: WAF rules, rate limiting, TLS settings, DNS controls, access logs, rollback planning, backups, release notes and enough monitoring to know when something is not right. None of that has to be sold as a horror story.
I like security work when it feels calm. Not complacent, just calm. The goal is not to make everyone anxious. The goal is to make the safer route the normal route.
Quiet controls are still controls. In fact, they are often the ones people actually keep using.
· Analytics and insight
Analytics should answer questions, not decorate reports.
I used to think analytics was mainly about having more numbers. Then you sit in enough conversations where everyone has a dashboard and nobody has an answer, and the problem becomes obvious.
Page views, sessions and traffic sources are useful, but they rarely explain the whole journey. For ecommerce work, the better questions are usually more practical. Where did people land? What did they see? Did they use search? Did they view the product? Did they add it to basket? Did the journey fall over before checkout?
That is why I care more about events and funnels than pretty reporting. A homepage banner is not just a banner if it is taking up prime space. A product card is not just a component if nobody clicks it. A search result is not a success if customers still cannot find the right item.
The trick is to track enough to learn, without turning the site into a surveillance project or creating data nobody will use. Good analytics should help teams make decisions. Keep this block. Rewrite that page. Improve that filter. Fix that step.
If a report cannot change what you do next, it is probably just decoration.
· Delivery and communication
The useful bit is often being the translator.
A lot of digital work gets stuck because everyone is trying to help from their own angle. The technical team can see the constraint. The commercial team can see the opportunity. The customer-facing team can see the pain. The problem is not that one group is wrong. It is that everyone is describing a different part of the same animal.
I have learned that one of the most useful jobs is translation. Not watering things down, and not turning technical work into vague business language. Proper translation means carrying the important detail across without losing the point.
A developer saying, "that is risky" might mean a data migration needs more testing, or that a shortcut will make future change harder. A stakeholder saying, "we just need it live" might mean a customer journey is currently costing money, trust or time. Both are valid. Both need unpacking.
The good middle ground is not compromise for the sake of peace. It is clarity. What are we trying to improve? What breaks if we rush? What can wait? What decision do we actually need today?
When that translation works, teams stop arguing around the edges and start making better calls. That is usually where delivery starts to move.
· Focus and balance
Learning to switch off is still work. Just a different kind.
At the start of 2025, I realised I had got quite good at staying switched on. Too good, really, for a normal life with early starts, commutes, family plans and evenings that should not still belong to work. Work followed me home in small ways: a bug I wanted to solve, a release risk I kept replaying, a message I should have worded better, or an idea that would not leave me alone.
Starting a fitness routine helped, but not because I suddenly became a different person. The useful bit was the boundary. For an hour or so, I had somewhere else to put my attention. Phone down. No tabs open. No half-reading messages while pretending to relax. Just the next set, the next walk, the next small bit of effort.
That changed more than I expected. It made consistency feel less dramatic. Turn up, do what you can, improve a little, then leave it alone until next time. That sounds obvious, but it was exactly the thing I was bad at with work. I could start. I could push. I could obsess. I was less good at stopping.
The lesson was not that fitness fixed everything. It was that having something outside work that asks for real attention can make work feel less like the whole room. I still care about doing good work. I just do not want every unfinished thought to follow me through the front door.