Surfon Articles

🧹 Why Your CRM Is Full of Dead Contacts, And What Proper Data Hygiene Actually Looks Like

Matt Wise — Wed, 13 May 2026 09:30:00 GMT

Every CRM I look at has a dirty secret. Somewhere between a third and a half of the contacts in it are dead. Not deleted. Not unsubscribed. Just inert. They haven't opened, clicked, replied, visited or bought in years.

And they're costing you money.

Most SME owners have no idea this is happening, because the CRM doesn't shout about it. The contact count keeps going up. The marketing team keeps adding to the list. The dead weight quietly accumulates underneath, like silt building up in a riverbed.

💸 The Cost of Carrying Dead Weight

Most SMEs don't realise that dead contacts have a real, measurable cost.

You're paying for them. HubSpot, Salesforce, Dotdigital and most others charge by contact count. Carry 50,000 dead contacts on a marketing platform that charges per contact and you're paying real money every month for storage you don't need. I've seen accounts where 60% of the contact bill was dead weight.

They're hurting your deliverability. Sending to people who never engage tells inbox providers your list is low quality. That damages delivery for everyone else on your list. The active subscribers you actually want to reach are getting fewer of your emails because the dead ones keep dragging the engagement average down.

They're skewing your reporting. Open rate of 12% sounds bad. But if half your list is dead, your open rate among engaged contacts might be 24%. You're making decisions based on a denominator that's wrong. You think your campaigns are failing when actually they're working perfectly well, just to a smaller-than-expected audience.

They're masking your real customer base. If you don't know who's actually engaged, you cannot segment, personalise or target effectively. Lifecycle campaigns rely on knowing who's active. Dead contacts pollute every behavioural segment you try to build, because the system doesn't know they're not really there anymore.

🔬 What Proper Hygiene Actually Looks Like

This is the routine I run for clients. It's not complicated. It just has to be done with intent.

1. Define what "engaged" means for your business. For most SMEs, that's opened, clicked or visited in the last 12 months. For higher-consideration B2B (think professional services or capital purchases), you might extend that to 18 or 24 months. The exact number matters less than having an agreed definition that everyone uses.

2. Tag everything by engagement status. Active, dormant, lapsed, dead. Make it a visible field on every contact, automatically updated. You should be able to filter to "all dead contacts" or "all active contacts" with one click.

3. Run a reactivation campaign on dormant contacts. One last attempt to re-engage. Make it good. A short sequence, two or three emails, with a clear question: do you still want to hear from us? Make the value clear and the unsubscribe easy. The ones who respond are gold. The ones who don't were never coming back anyway.

4. Suppress, don't delete. Move dead contacts to a suppressed list. Don't bin them. You may want them later for compliance or reference, especially under GDPR where you may need to demonstrate consent history. Suppression keeps them out of your active marketing while preserving the record.

5. Build hygiene into your routine. Quarterly audits. New contacts get tagged from day one. Bounces get processed weekly. Hard bounces get suppressed automatically. This becomes a discipline, not a one-off project. Otherwise the silt builds up again within 18 months.

🚨 The Mistake Almost Everyone Makes

The instinct most people have, when I show them their dead contact data, is to keep them just in case.

Just in case they come back. Just in case we need to reach them. Just in case there's a campaign down the line where they might convert.

I understand the instinct. But it's the wrong call. Carrying dead contacts is like carrying old stock. It costs you money, it confuses your reporting, and it makes you look like a worse marketer than you actually are. The 5% reactivation rate from a hypothetical future campaign does not justify the ongoing damage to deliverability and reporting clarity.

When I cleaned up the database at Goodwood, we cut email volume by 60% while increasing engagement, partly through better targeting and partly through ruthless hygiene. Same principle applies in a business of any size.

⚖️ A Quick Word on GDPR

Cleaning your list isn't just commercially sensible. It's also part of your obligations under GDPR.

Holding personal data indefinitely "just in case" isn't compliant. The principle of storage limitation says you should keep data only as long as you have a legitimate reason to. A contact who hasn't engaged in three years is a hard case to defend if you're audited.

Suppression with eventual deletion (after a defined retention period, documented in your privacy policy) is the cleaner answer. It protects you commercially and legally at the same time.

🧭 Final Thought

Your CRM should reflect the reality of your audience. Not the fiction of who they were three years ago.

If you've never run a proper hygiene exercise, the first one will probably reduce your list by 30 to 50%. That feels frightening. It shouldn't. The 50% you keep is the half that was actually paying attention. Everything else was noise.

When was the last time your CRM had a proper clean? If you can't remember, that's the answer.

📊 The Five Numbers Every SME Owner Should Look At Weekly (And the Ten You Can Ignore)

Matt Wise — Mon, 11 May 2026 11:14:59 GMT

The dashboards I see in most SMEs are full of widgets nobody understands.

Bounce rate. Average session duration. Page views per visit. Engagement rate. Reach. Impressions. Click-through rate on emails sent six weeks ago. Branded search volume. Domain authority. Social share of voice.

Lots of numbers. Almost none of them useful.

If you run a small or mid-sized business, you don't need fifteen metrics. You need five. Here's what they are, and why almost nothing else really matters at the leadership level.

✅ The Five That Actually Matter

1. Weekly qualified leads. Not all enquiries. The ones that fit your ideal customer profile. If this number isn't moving, nothing else matters. Calculate it by counting only the leads that hit your basic qualification criteria (right industry, right size, right role, real intent). A weekly view smooths out daily noise and shows whether your demand engine is actually working.

2. Lead-to-customer conversion rate. What percentage of qualified leads actually become paying customers. This tells you whether the issue is in marketing or sales. A low number with high lead volume means sales is the problem. High conversion with low volume means marketing is the problem. Most SMEs cannot tell which it is because they don't measure both.

3. Average customer value (first 12 months). What a new customer is actually worth in their first year. Without this, you cannot work out what you can afford to spend on acquisition. Almost every SME I work with knows their average order value but not their first-year customer value, which is a much more useful number.

4. Customer acquisition cost. What it costs you, all in, to win a new customer. Marketing spend plus sales time plus tooling, divided by new customers won. Combined with average customer value, this gives you the only marketing question that actually matters: are we spending less to acquire a customer than we make from them, and by how much?

5. Repeat purchase or retention rate. Whether your existing customers come back. The single biggest lever in most SMEs and the one almost nobody is measuring. It's much cheaper to grow revenue from existing customers than to acquire new ones, but you can only do that deliberately if you're tracking the number.

That's it. Five numbers. You can put them on one A4 page and have them ready every Monday morning.

🚫 The Ten You Can Mostly Ignore

Page views (without context, this tells you nothing about commercial outcomes)
Time on site (a high number can mean confusion as easily as engagement)
Bounce rate (changed definition in GA4 anyway, often misread by people who learnt it on Universal Analytics)
Email open rates (broken by Apple Mail Privacy Protection since 2021, no longer reliable)
Followers and likes (vanity unless tied directly to revenue)
Reach and impressions (same problem, social media's favourite empty calorie metrics)
Newsletter sign-ups (only useful if they convert downstream, which most don't)
Form submissions (only useful if qualified, and most aren't)
Cost per click (only useful if tied to CPA and CAC, otherwise a meaningless input metric)
Brand impressions (almost never tied to real outcomes in an SME with a small budget)

None of these are bad. They're just not what you should be running your business by. They're operational metrics, not commercial ones. Useful for the person running the channel. Not useful for the person running the business.

🧠 Why This Goes Wrong

Most SMEs adopt the metrics their tools surface by default. Google Analytics shows you bounce rate, so you measure bounce rate. The agency report includes reach, so you watch reach. Mailchimp shows open rate prominently, so that's what gets reported.

The result is dashboards that feel busy but tell you nothing about whether the business is actually growing.

The five numbers above are harder to capture, because they require connecting marketing data to sales data to finance data. That's exactly why so few businesses do it. And exactly why doing it gives you an edge.

When I was at The Jockey Club running CRO across a £40m+ ecommerce estate, the move that actually shifted commercial conversations wasn't a clever new dashboard. It was getting everyone in the leadership team looking at the same five numbers each week, with the same definitions, in the same format. Once that's in place, decisions get faster and arguments about "what's really happening" stop.

🛠️ How to Actually Build This

If your reporting is currently the wrong shape, here's how to fix it without spending months on it.

Define each of the five metrics clearly in plain English. Write the definitions down. Get the leadership team to agree.
Identify where each number lives. Some will be in your CRM, some in your finance system, some in your analytics tool.
Build the simplest possible weekly export, even if it's a manual job at first. Don't wait for the perfect dashboard.
Review them every Monday for at least eight weeks before changing anything. You need to see the patterns.
Automate the export only once you're sure the metrics are right. Building a dashboard around the wrong numbers is wasted work.

🧭 Final Thought

If your weekly leadership conversation is about reach and impressions, you've got a measurement problem.

The owners I work with who have the clearest picture of their business are the ones who've stripped their reporting back, not added to it. Five numbers, looked at every Monday, beats fifty numbers looked at by nobody.

What are the five numbers running your business right now? If you can't list them in 30 seconds, that's the gap.

🤖 AI in Your CRM: Where It Actually Saves Time, And Where It Just Creates More Noise

Matt Wise — Thu, 07 May 2026 09:45:00 GMT

Every CRM platform has gone all in on AI in the last 18 months. HubSpot has Breeze. Salesforce has Agentforce. Zoho has Zia. Pipedrive has its own AI assistant. The marketing on all of them sounds roughly the same: smart automation, predictive insight, agentic everything.

The reality, when you actually use these features in an SME, is far more uneven.

Some are quietly brilliant. Some are demo-friendly and useless. And some actively make your team's job harder while charging you a premium for the privilege.

Here's what I've found after rolling these out and ripping them out across multiple SME accounts.

🟢 Where AI Genuinely Earns Its Keep

After running these tools across multiple client accounts, here's where I've consistently seen real value.

Data enrichment and deduplication. AI is genuinely better than humans at spotting that "Sara Smith" and "Sarah Smith" at the same company are the same person, especially when you scale to thousands of records. This alone can save a CRM admin a day a week. When I was setting up the unified CRM at Goodwood, this kind of pattern matching was the difference between a clean database and a mess of half-duplicates that nobody trusted.

Email and meeting summaries. Auto-generated notes that capture what was actually agreed in a sales call, populated back into the deal record. Useful, time-saving, low risk. The salesperson reviews it, edits anything wrong, and saves twenty minutes per call. Across a team of three, that's a couple of working days a week reclaimed.

Forecasting and pipeline scoring. Not magic, but the pattern recognition is good enough to flag deals that look stuck or contacts who've gone cold. A useful prompt for sales managers in the Monday review. Not a replacement for judgement, but a sensible filter.

Content drafting for sequences. Not the final copy. But a starting point that saves 20 minutes per email is genuinely valuable. The trick is treating it as a first draft, not a final output. Marketers who use it that way get value. Marketers who use it to skip the writing process entirely get the bland, generic AI prose everyone has learned to ignore.

🔴 Where AI Is Mostly Theatre

Auto-generated reports nobody reads. "Here's a summary of your week" emails that get ignored after the first three. The format never quite matches the question the user has, and the insights are generic enough to be useless. These are the AI features the platform demos love but the customer ignores within a month.

Chatbots pretending to be salespeople. Customers can spot these in seconds, and the conversion data shows it. There are good use cases for chatbots (FAQs, simple routing, account self-service). Pretending to be a human salesperson is not one of them. It damages trust before the conversation has even started.

Predictive lead scoring with no historical data. If you've got 200 leads a year, the model has nothing to learn from. The "AI score" is essentially random. SaaS platforms will happily turn it on for you anyway, because it makes the demo look impressive.

Generative AI writing your customer comms unsupervised. I've seen this go wrong in expensive ways. AI doesn't know your tone, your context, or what you promised the customer last week. The horror stories I've heard from my network involve AI-drafted apologies for the wrong issue, AI-generated price quotes that bear no relationship to the real pricing, and AI customer service responses that quietly accept liability the business never agreed to.

🧠 The Real Question Nobody Asks

Most SMEs are buying AI features without first asking the boring question: what slow, repetitive task in our business actually deserves automation?

The answer is usually not "write our marketing emails". It's usually something like "match incoming form submissions to existing contacts" or "flag deals that haven't moved in 14 days" or "summarise this week's customer support tickets into a single readable digest". Less glamorous. Far more useful.

If you start with the task and work backwards to the tool, AI can be transformative. If you start with the tool and look for tasks to apply it to, you end up with a list of features running in the background that everyone has quietly switched off.

🧰 A Practical Decision Framework

Before you turn on any AI feature in your CRM, run it through these four questions:

What specific task does this replace, and how long does that task currently take?
If this works, what would my team stop doing?
What's the worst case if it gets it wrong, and is there a human in the loop to catch that?
How will I measure whether it's actually working in 90 days?

If you can't answer all four, leave the feature switched off. The AI features that survive contact with a real business are the ones you can defend with these answers. The ones that fail tend to be the ones where the answers are vague.

🧭 Final Thought

AI in your CRM should be invisible. If you can feel it, it's probably annoying you.

The good implementations free up time you didn't know you were losing. The bad ones add another layer of complexity to a system that was already too complex.

Before you turn on the next AI feature your CRM offers, ask: what would I stop doing if this worked? If you can't answer that, leave it switched off.

What AI features have actually made a difference in your business? And which ones are you quietly ignoring?

📉 Why Your Email Open Rates Dropped in 2025 (And What To Do About It)

Matt Wise — Wed, 06 May 2026 08:29:31 GMT

Sometime in late 2024, a lot of SME marketers noticed something strange. Open rates started dropping. Click rates dropped with them. Nothing in the campaigns had really changed.

Then 2025 hit, and the trend got worse. Not slightly worse. Materially worse. The kind of drop where the marketing team starts having uncomfortable conversations with the commercial director.

If your numbers fell off a cliff and you've been blaming your subject lines, you can stop. The subject lines are probably fine.

📬 What Actually Changed

In February 2024, Google and Yahoo brought in new bulk sender rules. They were quiet about it at first, but the practical effect was huge.

If you send more than 5,000 emails a day to Gmail addresses (which most B2C SMEs do, and a surprising number of B2Bs do too), you now have to:

Authenticate properly with SPF, DKIM and DMARC
Keep your spam complaint rate under 0.3%
Make unsubscribe a one-click action, with the link processed within two days
Maintain genuinely engaged lists, not just lists you've built up over time
Senders who haven't done list hygiene in two years getting throttled hard
DMARC set to "none" because nobody knew what to set it to, or set up wrong because the IT team did it without talking to marketing
Welcome emails landing in spam because the sending domain has no warmed-up reputation
Engagement-based segments not being used, so cold subscribers drag down the deliverability for the warm ones
ESPs sending from shared IPs that have been blacklisted because of someone else's dodgy practices

The crucial bit is the engagement piece. Gmail and Yahoo are both now leaning much harder on engagement signals to decide where your emails go. Inbox or promotions tab. Promotions tab or spam. Spam or never delivered at all.

Nobody is going to tell you that this is happening. There's no warning email. You just notice that the open rate that used to be 24% is now 11%, and the click rate has dropped with it.

🧪 What I'm Seeing in the Wild

Across the audits I've run since the start of 2025, the same picture keeps appearing.

The interesting bit is that the businesses doing this badly aren't doing anything new. They're sending the same campaigns they sent in 2022. Same list. Same content. Same cadence. The world moved, and they didn't.

When I was at Goodwood, we cut email volume by 60% while increasing engagement, just by getting smarter about who got what and when. That same logic applies now, but with more urgency. Frequency without engagement is no longer just inefficient. It's actively damaging.

✅ What To Actually Do

1. Check your DMARC record. If it's set to "none" or you haven't got one, that's job one. Set up SPF, DKIM and DMARC properly. Move DMARC from "none" to "quarantine" once you're confident the legitimate mail is authenticating. There are free tools (mxtoolbox, dmarcian) that will tell you exactly where you stand in 30 seconds.

2. Clean your list. Anyone who hasn't opened in 12 months is probably hurting your sender reputation. Suppress them, or run a reactivation campaign and then suppress those who don't respond. This will feel scary because you'll be removing thousands of names. The names that stay will engage at 2 to 3 times the rate of the average you had before. Your numbers will look better and so will your deliverability.

3. Segment by engagement. Send your most active people more, your least active people less. Stop blasting everyone the same thing. The simplest split: active (engaged in 30 days), warm (engaged in 31 to 90 days), and cold (no engagement in 90+ days). Send to actives often. Send to warm carefully. Reactivate cold or suppress them.

4. Warm new sending domains slowly. If you've migrated to a new ESP or domain, ramp up gradually over four to six weeks. Otherwise you'll torch your reputation in a fortnight and spend six months trying to recover it. This is the single most common mistake I see when companies switch platforms.

5. Look at where you're landing. Use a tool like Mail Tester or GlockApps to see if you're hitting inbox or promotions across major providers. Most SMEs have never checked. You can spend £20 once a quarter and have a much clearer picture of what's actually happening to your emails after they leave your ESP.

🔭 How to Monitor From Now On

Email deliverability is not a fix-once project. It's an ongoing monitoring discipline.

The simplest version: every quarter, check your DMARC reports (your ESP or a tool like Postmark's free DMARC monitoring will surface them), run a deliverability test through Mail Tester or GlockApps, and review your engagement segments. If active subscribers are dropping or cold subscribers are growing, something needs attention.

Most SMEs will never put this discipline in place. The ones that do quietly outperform their competitors on email by a wide margin.

🧭 Final Thought

Email is still the highest ROI channel for almost every SME I've worked with. But the rules of the game have changed, and the businesses still treating it like 2018 are quietly bleeding revenue.

It's not glamorous work. Setting up DMARC isn't going to win anyone an award. But it might be the difference between a 22% open rate and a 9% one, and that gap is where your email programme either pays for itself ten times over or quietly dies in the promotions tab.

When was the last time you properly audited your sender reputation? If the honest answer is "never", that's where I'd start.

🛠️ The 7 HubSpot Mistakes I Find in Almost Every SME Account I Audit

Matt Wise — Tue, 05 May 2026 11:27:46 GMT

Every HubSpot account I open looks different on the surface. The dashboards are different colours, the pipelines have different names, the contacts are in different states of disrepair.

But the mistakes? The mistakes are nearly always the same seven.

I've audited a lot of HubSpot accounts now, across very different businesses, and I've started keeping a tally. Some are scaled-up service businesses doing seven-figure revenue. Some are early-stage startups still finding their feet. The pattern is the same. Different industries, different price points, same broken plumbing underneath.

Here's what keeps coming up.

🔍 The Pattern I Keep Seeing

When SMEs buy HubSpot, they usually do it for one of three reasons. They've outgrown a spreadsheet. They've been told they "need a CRM". Or someone sold them on the idea that automation would magically fix their pipeline.

What happens next is predictable. The trial gets set up by whoever happens to be free that week. A few fields get added. A pipeline gets sketched out. Maybe an automation or two gets switched on because someone watched a HubSpot Academy video on a Friday afternoon. And then it stops.

Eighteen months later, I get the call. Usually it's because reporting has stopped making sense, or because the sales team have lost faith in the pipeline, or because email performance has fallen off a cliff. The symptoms vary. The root cause is almost always the same set of foundational mistakes.

🚩 The 7 Mistakes (Roughly in Order of How Painful They Are)

1. Lifecycle stages set up backwards. People mark contacts as "Customer" the moment they enquire, which means every report on conversion is broken from day one. I've seen accounts where 80% of the database is tagged as "Customer" and the actual paying customer count is a tenth of that. Every funnel report that exec team looks at is meaningless. Nobody notices because the numbers look big.

2. No lead source tracking. Without it, you cannot tell which marketing channels are actually working. Most accounts have this field, but it's empty or set to "Other" on 80% of records. The result is you've spent five years building up a CRM and you still cannot answer the most basic marketing question: where did this customer come from?

3. Duplicate contacts everywhere. Same person, three records, all with slightly different data. Different email addresses, slightly different name spellings, different phone formats. Nobody owns the cleanup. Every email send goes out two or three times to the same person. Every report double-counts. Trust in the data quietly evaporates.

4. Automation loops that nobody's checked in months. Workflows still firing for products that no longer exist. Sequences still emailing people who unsubscribed two years ago because someone built a workflow that bypassed the unsubscribe logic. Welcome flows triggering for existing customers because the entry criteria were never tightened up. This is the one that genuinely embarrasses people when I show them what's running.

5. Pipelines built for the salesperson, not the customer. Stages like "Hot lead" and "Warm lead" that mean different things to different people, with no exit criteria. A deal sits in "Negotiation" for six months and nobody can say what would actually move it to "Closed Won". Pipelines should be built around the customer's buying process, not the seller's mood.

6. Forms that don't talk to the CRM properly. Submissions land, but the data doesn't map to the right properties. Job title goes into a generic notes field. Industry isn't captured at all. The form looks fine to the customer, but everything downstream is broken. Segmentation becomes impossible because the data isn't where it needs to be.

7. No reporting structure. Dashboards full of widgets that nobody looks at, while the three numbers that actually matter aren't on a screen anywhere. Or worse, dashboards built by an admin two years ago that the leadership team have stopped trusting because the methodology was never documented.

💸 The Cost of Leaving These Alone

Most SMEs treat these mistakes as low priority. Things to sort out one day, when there's time. The truth is they're costing real money every month they're left.

Bad lifecycle data means marketing spend gets attributed to the wrong channels, so budgets shift in the wrong direction. Duplicate contacts inflate your HubSpot bill (you pay per marketable contact) and damage your sender reputation. Broken automation creates customer experience disasters that you find out about when someone forwards you the angry email.

When I worked at The Jockey Club running CRO across a £40m+ ecommerce estate, the difference between a clean data foundation and a messy one was measured in seven figures of revenue. The stakes are smaller in an SME, but proportionally the same. A 5% lift in lead quality, applied to a year of marketing spend, pays for the cleanup ten times over.

🧰 What a Fix Actually Looks Like

The good news is that none of this requires a six-month transformation programme. Most HubSpot accounts I work with can have these seven issues meaningfully addressed in a focused week or two of work.

The order I usually run it is:

Audit and document what's currently broken (the boring but essential first step)
Fix lifecycle stages and lead source first, because everything else depends on these
Run dedupe and basic data hygiene
Pause and review every active workflow, killing or fixing what's broken
Rebuild the pipeline with proper exit criteria
Map forms to properties properly and test end-to-end
Build a stripped-back reporting layer focused on the numbers that actually matter

That's a sequence, not a menu. Doing them out of order tends to mean redoing earlier steps.

🧭 Final Thought

If your HubSpot account has been running for more than a year and nobody's properly audited it, there will be at least three of these in there. Probably more.

The good news is that fixing them is rarely a six-month project. The bad news is that nobody's going to do it for you unless you make the time, or bring in someone whose job it is.

What does your HubSpot account look like under the bonnet? If you've inherited a setup that nobody trusts, get in touch. Happy to take a look.

AI is the New Machine Learning: Beyond the Buzzword

Matt Wise — Wed, 28 Jan 2026 15:06:46 GMT

AI is everywhere. Everyone’s talking about it. Every other meeting, someone’s saying "AI-powered this" or "AI-driven that." Sound familiar? It should—because we’ve seen this before.

I remember when machine learning became the hot buzzword. Businesses rushed to jump on board without really knowing what it meant or how to use it properly. Now, AI is at risk of falling into the same trap.

Why I Think This Matters

AI has the power to transform businesses—but only if we understand what it can (and can’t) do. It’s not some magic solution that can fix things overnight. Without solid data, clear processes, and a defined strategy, it’s just another expensive trend that doesn’t deliver.

The AI Buzz: A Familiar Pattern

It's happened before and it will again — leaders and businesses get swept up in the excitement of shiny new tech. First it was machine learning, then big data, then IoT. These were all legitimate innovations, but often embraced without ever being fully realised.

Now, it’s AI’s turn. Everyone wants to say they’re using it—but too often, they’re unclear on what problem it’s meant to solve.

The problem with buzzwords is they oversimplify complex systems. AI isn’t about plugging in a chatbot or getting a quick sales forecast. It takes real groundwork—things like ensuring your data is ready, your platforms are integrated, and your teams know how to act on the insights AI provides.

What Are Some Things AI Can Do Well

When it’s done right, AI can:

Automate repetitive tasks: Think admin work, reporting, or sending follow-up emails—freeing teams up for more valuable work.
Spot patterns and trends: AI can flag customer behaviour, sales trends, or risks before anyone else even notices them (if your data is robust).
Deliver personalisation at scale: Tailored emails, product recommendations, and more—without massive manual effort.

But AI is only as good as the systems and data that support it. And this is where a lot of businesses stumble.

AI is Not a Fix-All

Let’s cut through the hype—here are the key things to keep in mind:

Data quality is key: Messy data leads to messy outcomes.
AI doesn’t replace strategy: Insight still needs human judgement.
Integration is critical: AI should feed existing workflows, not sit in isolation.
Beware of ethical risks: Bias, compliance, and data protection matter.
Cost vs. value: Sometimes simpler systems are the smarter choice.

My Approach to Embracing AI — Smart, Not Rushed

AI is only going to get better—and businesses that avoid it might fall behind. But that doesn’t mean rushing into AI projects just to look innovative.

Define the problem before choosing the tool.
Make sure data, processes, and teams are ready.
Focus on small, practical wins.

Final Thoughts

AI has incredible potential—but only if we move past the hype and focus on outcomes.

What do we actually need AI to do?
Are we ready for it?

What’s been your experience with AI? Has it changed how your business works—or is it still more talk than action?

💳 From Checkout Tricks to Coachella: When CRO Crosses the Line

Matt Wise — Wed, 28 Jan 2026 15:02:47 GMT

More than half the tickets were financed.

Roughly 60% of attendees used Buy Now Pay Later (BNPL) tools—installment schemes dressed up as convenience—to secure a spot. Some paid $49.99 up front, followed by months of payments and “service fees” that quietly padded the final cost.

It’s easy to laugh at the absurdity of financing a festival ticket. But honestly? This is just the public-facing version of something I’ve seen quietly spreading through CRO teams for years.

🧪 When We Tested Klarna

In my previous role, we ran a test offering Klarna at checkout for premium race day and hospitality bookings. We wanted to see if giving customers the option to split payments would drive upgrades.

Technically, it “worked.” Some segments spent more. Others didn’t. A few underperformed. But the test revealed something more important than any uplift:

It made me deeply uncomfortable.

Because here’s the thing: most people aren’t Klarna-ing because it’s convenient. They’re Klarna-ing because they don’t have the money.

🧠 CRO Has an Ethics Problem

We often treat CRO like a maths problem: run test A, measure conversion B, deploy the winner.

But optimisation without guardrails becomes exploitation.

Countdown timers that aren’t real
Financing options pushed before total cost is even visible
Hidden fees made to look like service upgrades
“Nudges” that cross the line into manipulation

And now, leisure is being sold on credit. Coachella, Lollapalooza, gigs, theatre tickets—people are financing joy in 6–12 month instalments. Because it’s marketed as freedom.

But it isn’t. It’s dependency.

📉 The Tipping Point

According to the FT, Klarna reported a 17% year-on-year increase in defaults in Q1 2025 — $136 million in unpaid loans. More customers are failing to repay what was pitched as “zero-risk” money.

So what happens when BNPL becomes the norm for everything from concerts to groceries?

We normalise phantom affordability. We trade long-term trust for short-term conversion.

And worst of all? We congratulate ourselves when the CRO dashboard turns green.

💡 What’s Our Responsibility?

As digital leaders, we have to ask harder questions.

Not just: Did this improve conversion?

But:
– Would I offer this to someone I care about?
– Does this enable better choices—or just more spend?
– Are we removing friction, or removing self-control?

Sometimes the ethical red flags are obvious. Other times they’re hidden behind slick UX and softened by a “user-first” narrative.

But make no mistake: the more we rely on financing to sell experiences, the further we drift from sustainable, customer-centric growth.

🧭 Final Thought

Coachella didn’t fail to sell out because people lost interest. It failed because people couldn’t afford it upfront — and the industry’s response was to make it easier to go into debt.

That’s not innovation. That’s abdication.

Just because it converts, does that make it right?

If you’ve seen (or run) CRO tests that made you wince, let’s talk. I’m collecting ethically grey tactics we’ve all encountered — because owning that discomfort is the first step to doing better.