A few days to think, and now a reflection on MACH in Amsterdam for those who are interested.

I’ve said this before, but the MACH Alliance has nicely pulled off a pivot towards the following - “You will have a better AI outcome if you use a composable stack”. I suspect the VTEX team are a little irked, but they should definitely not have stormed off IMHO.

Anyway, those days of didactic stack religion are over, and there is now a different atmosphere at these events, bullets to follow:

• A little more humility from members/vendors. Less charging aroud the venue looking for people to sell to who match their ICP.

• More convivial conversations across roundtables and such.

• A smaller role for the original MACH overlords, and much more from the Ambassadors and customer speakers

• At least one award winner was a Shopify user - given an award for the way they were using Patchworks through Claude to integrate and operate 12 solutions through a single interface… a new world indeed.

We’re in this together

There is, without doubt, the sense of a shared challenge. AI is (slowly) offering a different view of the future to everyone.

Where anyone can build a SaaS solution, the roles of vendors and customers really changing.

A “heard” version of “new” seems to be -

1. Customer creates “vibe coded” POC…

2. Vendor (partner agency or trusted SaaS) examines and offers to add to roadmap, or builds for client

3. Or - Customer proceeds to build and launch production-grade tooling, and agency maintains 9or customer team)

4. Less often, vendor builds new feature and releases for small batch of test clients who have asked

But these are uncertain times. This event was VERY well attended - about 81% rate of people turning up who registered. Considering the NRF paris disaster, Human X’s empty room in Amsterdam the week before, and some questionable numbers in the U.S. at Shoptalk - why is this?

Why did people turn up?

The message that “we’re in this together” seemed to really resonate. Nobody has the answer so let’s see if we can make progress through discussion - I think that honesty is one reason.

I also saw vendors in a lot of sessions. Listening and contributing - their views are really valid if they’re honest, and many were. A good number of people mentioned how much they enjoyed the dialogue on offer - more than the presentations in some ways.

I met a 5 agencies who claimed to have grown in the last 12 months. Maybe their death knell has been predicted too early?

From London this event was unrecognisable to be honest - a fully recovered and renewed ecosystem perhaps?

Slow down FFS

What I’ve always enjoyed about the process of putting an Agenda together is realising that customers broadly agree on where the problems are, and how urgent they seem. On stage there is sometime a critic who says that speakers repeat one another, but I’d argue that’s because what their saying is true in the Enterprise.

On this occasion the following were “true”:

• There’s no Agentic traffic yet, so it’s next year’s problem and nobody is worried. Composable stacks are designed to add this channel easily.

• AI is both everywhere and nowhere. Content and data enrichment (Y), Instant checkout (Y), eComm (N), Redundancies (N),

• Discovery and search are more urgent. The older SaaS is already unsuitable and many are choosing to invest in a build of their own, making discovery part of their core IP.

• There were three companies I chatted to who are weaning themselves off the Google/Meta tax. This in turn affects their their overall profit, their ability to invest, their NPS scores and repeat purchase rates. All were D2C pureplays

• Amsterdam is beautiful and everyone wants to visit more often :)

A few other comments if you’ll forgive me.

It drives me mad when salespeople do not attend the sessions, and I did meet a group of 5 member salespeople who spent the morning chatting to one another rather than doing this. I thought we’d learned that lesson.

I doubt we’ll run a “governance and security” track in the next MACH X in Boston, but the European Psyche enjoyed it enormously.

That’s all - #norocketships main issue goes out tomorrow as scheduled.

Jamie

Before I begin. It's worth saying that there is no LLM here, in fact I slightly rejoice in never going very much back over anything I write… so expect accidents and evidence of no proofreading 🙏.

The future of Commerce sept 2027

This is a love letter to digital, a world in which I have worked for nearly 20 years.

Whilst the digital world has evolved significantly over 20 years, most of the focus has been on creating websites that were able to transact safely, securely and on brand - and then subsequently finding audiences beyond the initial enthused early adopters of the internet.

In 2026 this channel is under stress as never before, and something needs to give in my humble…

As I see it, growth online is becoming almost unaffordable using the methods most deploy. New customer acquisition online is becoming almost completely unprofitable, as paid costs are driven up, inboxes are fragmented and “organic” across the board somehow unpopular and out of fashion. Of course, our Californian overlords are currently creating a new jigsaw for digital with ever smaller pieces, some with nothing on them, and some that are just missing - on purpose.

It’s worth noting my credentials here. As well as moderating a group of 300 brand-side customers, I speak to a good number of leaders across operations, owners, digital and tech in designing conference Agendas - without an agenda of my own I might add. These people know that this is happening. They stopped believing as an organisation that digital was the “engine of growth” for their business a while ago, and whilst it’s not said in public yet… they’re looking elsewhere.

Somehow, online has become wrapped up in its own ecosystem so that (for example) paid media managers measure themselves based on the size of their budget, the new shiny tech toy is more appealing than extracting value from the current stack, and (of course) that data cleansing project is still on the back burner because…. It’s hard and unsexy.

Coincidentally, stores have been stuck in the equivalent of this same rut for decades - change was resisted, hierarchies persisted and a slow decline was envisioned.

In my view we find ourselves at a crossroads of sorts, where it seems as if online is genuinely under threat, store retail is recovering, and big tech appears determined to decimate digital by rendering it impossibly unprofitable.

There are outliers of course - I could point to Flaconi and Pott’d and many more who are spending way less than 10% of revenue on paid to support their digital channels, and who are growing really quickly and profitably… but they are few and far between.

For years I have warned software vendors against telling retailers how to run their business and so I will shy away from giving out hordes of advice but there are some bullets that I would offer as good counsel that people really should start thinking about before the worst that can happen happens:

1 - Take a really good hard look at the way in which you spend your Media money. I know several very successful fast-growing businesses around Europe who do not pay the Google or meta tax in huge order, but instead focus on brand, service, retention and community - and they do not avoid the hard things.

2 - Examine the team motivation that you have that potentially causes you to have reliance on (or addiction to) that paid media spend. Are you spending it because everybody just thinks you should have to - we all know that there is a collective momentum that forces this spend: The agency; Google / Meta themselves; Shonky Analytics - all of these things are telling you to spend more money on paid because it's supposed to be very efficient but really and truly is it are you just a victim of this reinforced belief?

3 - Ensure that you give every possible channel the right amount of examination before you settle on an online only advertising program. There are significant opportunities in physical retail and as we all know there are significant opportunities in wholesale. It's no secret that Represent the clothing business has created a great deal of its growth over the last two or three years from stores and wholesale as a set of new channels. I could cite all sorts of examples of retailers who have managed to create a multi-channel strategy that has really paid off, growing from D2C.

4 - Whichever way you end up investing you will need to start using online as much more of a customer retention and growth channel than a customer winning channel just because of the cost issue. which means that organize enriching your customer data, creating the right kind of cohorts and communications that support stickiness and repeat purchase is is where a lot of your online Investments should be going. People at ny conferences have been saying it for years but in general most brands have not prioritized it because the lure of cheap social has been so shiny that it's been very hard to say no to that.

5 - Take a really good hard look at that thing that you call a Loyalty Program. does it incentivise the second, third and fourth purchase from a customer, and if it does not and if the incentives are not what they should be think really carefully and make the Adjustment now. Loyalty is a luxury that online cannot afford in the current climate - a good strong inventive program is.

OK - enough doom and gloom, but this was as much for my own sanity as your benefit - thanks for your indulgence.

We’ll be back in a few weeks

Jamie

Vile Rumour & Gossip

• POC does not always reflect the platform delivered - No names on this one, but an eCom retailer told me this week that the POC tech they used for a new eCom platform was not the tech supplied once the contract was signed. Broken API’s, missing functions and inconfigurable front ends. They were forced to go back to the platform who lost 6 months in and change. Who lies like that?

• Do not build an AI team - We know some brand-side peeps who have these units, but a little like “digital” who kept the door locked a few years back, these don’t work. In most cases they’re building things people don’t want, and are already a laughing stock.

• The trouble with LinkedIn eating itself is... Over 20yrs we’ve all built a habit - 5 minutes 35 times each day, the drumbeat of biz dev, partnerships, event FOMO and competitor jealousy - Gone (almost over a single summer). I am seeing only “Suggested posts” and worryingly right wing bombast in my feed. Sad times. The alternatives are (frankly) a bit crap. Our CF community is very active in a way LinkedIn used to be, it seems odd that these platforms are driving us away don’t you think?

• Cimulate - overpromising sadly Salesforce bought these guys back in February, and SF’s many sales teams are out there doing demos. We happen to know that the business has left a trail of less than satisfied clients behind them (basically having sold an unfinished product), which given the scale of the sales & marketing machine now behind them is a worry. Please - sort out the clients first - DM us and we’ll give you a list :)

Outreach isn’t dead

I hate the term “reaching out” (No, I’m not in the Four Tops so it’s not allowed), but so many people come to me nowadays and ask about SDR tactics that I’ve decided to write about it a little.

Firstly, content and personalised DM is the way forward for the first contact. Do not even THINK about making a phone call until you’ve done a great job at this - it’s called marketing and other industries outside of tech have been doing it for years.

I’ve watched a two person team (founder of Slip Tash Grossman and her marketing mgr) create amazingly creative DM that supported their networking activities and secured the first meeting almost without a call - but these DM pieces take time - a day at least per prospect. Are you willing to do that work? In the absence of a team and a budget then Tash had no choice - I wonder if she’d do the same now that Slip has grown?

It’s also good to hear that some really modern tooling has emerged from the bloodbath of things like Apollo - Ample Market is being adopted rapidly because unlike so many others it contains some good reliable intent data, coupled with the lists you need. SDR’s then do not annoy humans with no interest at all :)

If you’d like to see any examples of really good creative B2B DM then message me and we can see if we can share.

Client-side fatigue

The summer has been a tough one for most brand teams. The year hasn’t turned out to be an easy one to move jobs, and so a lot of people are 4-5yrs in - and sad they’ve not had the bump or the exit they wanted.

Many of those in senior eCom & Digital roles feel they’re a resource rather than a leader, the word “firefighting” comes up a great deal - especially over the summer as teams prep for peak with fewer team members.

Why is this important? It plays to the commercial environment that buyers and sellers are operating in. 5 years in a business and you have both delivered and lived with the things you’ve sent live - for better or worse. Your CEO has seen what you have, and your peers are aware of what you’re really like. You either now have autonomy (if you’re good) or a controlling CFO (if you’ve been bad).

The AI Estate Agents

Whenever I read a post about Anthropic investing in infrastructure, it makes me think of the ways Data Centres are typically funded. At the top end, they’re made to look like part of AI, but in reality this is just dirty old bare metal server stuff, with words made up to cover up the machines inside.

Data Centres are Real Estate deals, with people who invest in commercial property doing so because a DC offers a high yielding investment. These guy aren’t looking into unused Nvidia compute, or even empty racks... and a lot of this debt is being hidden amongst small investors.

AI is dressed up and ready to float, but the story is a dirty one.

Another insult from our ecosystem, this time from Jasmin Guthman who sent us this issue’s insult. Who do you know that is a Cocktupus?

And finally - some events to think about

MACH X - Late Sept - I’m on Agenda duty so it’ll be epic of course

Customer X - Mid-Oct - Ian Jindal’s latest idea... should be excellent

Jamie

Vile Rumour & Gossip

• Salesforce are rearming - Always interesting to see who SF acquires as AI business models take shape. The SF walled garden now has M3ter (a very good usage billing platform), TIN (a behemoth AI agent for CS) and more as they take advantage of the lowering of prices in SaaS. Still - $3.6bn for TIN is a whole heap of cash. At least the US govt can’t turn it off when they feel like it right?

• Shopify’s international fudge - I’m often told I’m too harsh on vendors, but the Commerce Futures group is rife with no small frustration regarding “markets” v single storefronts. I get it - Shopify’s home market is so vast that the ROW is a distraction of sorts, but with so much developer talent it should be able to re-architect for international merchants properly without needing many instances. The European enterprise isn’t an option if this is where we stay, and for those who think me too harsh, the Shopify disciples would have you believe it’s capable of running a global business. Yes - if you have 36 storefronts :(

• The haggle is real - Renewals are getting very interesting. We’re hearing of Search businesses agreeing to 1yr extensions with huge discounts (and sighing with relief). Meanwhile Rezolve.ai is out there trying to acquire these businesses on the cheap. We’re hearing rumours that they’re trying to buy Athos... which would be sad given that out of the many mergers around the Athos org has integrated pretty well.

• The LLM cute name contest - I realise we’re all going to become conspiracy theorists as we age, but it’s not just me I promise. LLM’s are all being given “benign” names... taking words from the Tolkien and Greek myths to make them feel super-approachable. The irony that something called “Fable” is the subject global grumpiness is sad, but I do wish the vendors would stop making their models sound cute. They’re not cute, right?

Anti-Loyalty

I’ve tried saying it quietly, but nobody seems to listen. So - a little bit louder.

There are no loyalty programs, there are no loyalty vendors, there are only incentives.

Of course, we all have brand affinities. I’m loyal to Aston Villa, McNair shirts, Petite Peonies. I don’t have to be told that... I just am.

But the economics of incentives - there’s a thing of beautiful complexity, a thing of interest. That’s where “earn and burn” is weakest - tearing apart any affinity with the brand whilst pretending the opposite is in fact happening.

If AI gives us anything it’s the ability to personalise incentives at scale. Copywriting for each individual based upon their behaviour and purchases, discounts that feel random and “gifty” but are actually evenly spread across the customer base.

Come on people - let’s try harder... and stop calling them loyalty programs, they’re just not.

Sovereignty

Where to start. As I design the next MACH Agenda this has to be present. European businesses will not adopt the AI that the vendors need them to if the models risk being turned off.

Global politics are front and centre here, and there’s just no way a global LLM is going to offer any guarantee of availability in the current climate. In Canada a few weeks ago we heard lots of chatter about “Sovereign GPU’s”, which seems like an obvious priority, but the reality is every market will need a sovereign model selection, the world is simply too uncertain right now.

In the UK we’ve got nothing, we’ve sold our soul. the French alone have an LLM - Mistral (not ripe for use though). Nothing in any other sovereign European state, and no sign either.

The vision Tim Berners Lee purports of us each having our OWN sovereign wallet is a very long way away, and whilst the word “Sovereign” has all sorts of awful connotations to do with “Empire” and worse, in the new world it really does mean “mine own”.

We should all think much harder about this subject. Before jumping in to every new model for fun... just hold that thought for a second. Count to 10 - have a ponder.

The LLM Feed - customers in the dark

We are used to things moving faster now... and given the focus on LLM’s across vendors and clients we had expected that in 3 weeks between our last issue and now, we’d have found someone who really understands the product feed for AEO.

Not so... vendors do not seem to grasp the opportunity, and clients are standing around confused. Somebody, anybody, help :)

And finally - our insult of the week. Please feel free to use freely. Don’t forget to find us on substack... A great piece about AI careers there and more to come we promise.

Jamie

The image above has been freewheeling around my head for months now.

Is this where we’ll end up as “agents” remove our critical thinking? And Claude has told us we’re amazing once too often?

I don’t want to lose the little true cognition I have. And yes, thinking and chatting is the lifeblood of our working lives....

So - let’s get into it - my opinions follow...

Vile Rumour & Gossip

• Recharge bought Skio on the cheap - Skio were doing $33m ARR, and got bought for 3x revenue. 2 years ago they’d have gone for 10x at least. Recharge has become largely dominant in their sector (except for Shopify’s own limited subs service), and Skio were better at lots of the elements of the Recharge toolkit. Customers are a bit sad of course, they know how this ends (more on that later). SaaS valuations are really screwed right now, and what that means for pricing, salaries, all of it - we really don’t know yet.

• The murky beginnings of Rezolve.ai - If you’re in the right community groups (or on reddit) you’ll have read about the dark arts of Dan Wagner’s not so new startup. The beginnings are largely unimportant now though. Out of a SPAC they’ve acquired Crownpeak & Advanced Commerce among others, including three search tools and a bunch of customers. And - $750m burning a hole in their pocket to be spent quickly.
  Murky or not, these are real customers with SaaS that is deeply embedded (some have had Fredhopper for 15 years)... so keep a close eye people, the agentic promise is clearly fluff for now. Oh and yes i asked them for a chat - I am being ghosted.

• Klaviyo’s CRM carcrash - A few weeks ago the CF community witnessed a new AE at Klaviyo sending a “hello, can we catch up” email to 50 accounts who had mostly left their businesses. Oh, and it came out of Salesforce... :) Still funny 10 days later.

Adyen & Talon one - what does it say about SaaS?

This transaction was very early in the cycle that SaaS has known. Talon One has had real momentum for a while now, and hadn’t actually raised that much cash... so really were in control of their own destiny.

I wonder if this has happened now because the thinking is that valuations will only get worse? I can’t see that sales momentum is an issue, the team are excellent and the product positioning has been strong.

The deal presents opportunities for others though. This is Adyen’s biggest acquisition, and they won’t be able to resist fiddling with their new toy. Inevitably, Talon One’s momentum will slow, great staff will move on, and Adyen’s conservative/cautious culture will take over.

Who will benefit? the established “loyalty” players for sure... but there are some bootstrapped and clever types out there. I suspect there will be far more of these deals this year, leading back to suites of SaaS which may never really be bought again....

Are we witnessing the demise of pureplay D2C?

As I wrote a few weeks ago, many D2C companies are spending 20-35% of their total revenues on customer advertising. All with three channels.

In 2005 Heinz (one of the biggest spenders) were lashing out 4.5% of sales revenues on marketing. This has since grown to about 8%, but D2C will never be profitable if they continue to spend at 20%+.

But continue to spend they must for now, because whilst growth is hard to come by, pureplay D2C has to spend big just to stand still. Squeezed by rising ad costs, supply chains, geopolitics and product ranges that do not have repeat purchase built-in - we worry for the sector itself.

Anyone up for opening some shops?

How we love the lion’s den

I got to witness a delicious 10 minute f2f sales pitch & interrogation last week. A senior architect from The Gap and the founder of a server-side tagging business had a really intense chat where deep knowledge was tested, eye contact maintained, subject-matter expertise displayed by both - and ultimately the test was passed, and a next step agreed.

It’s sad in so many ways that all these relationships now start with a zoom call. Face to face is ultimately faster and much more effective, and I know both salespeople and clients would rather start do it in person.

Perhaps in an AI world buying and selling will be the last bastion of critical thinking before we all turn into blobs on lilos :)

Anyway - more focus on these for the rest of the year from me, and more events in the Autumn. Happy Q2 to all and we hope you get the commercial outcomes you deserve (and not a penny more). And - a shout out to Tomo for the support with insults... Temu is a doozy.

Jamie

You’ll have seen by now we’ve added a podcast to the #norocketships offering, and we reckon episode 2 is a doozy. The slightly reluctant interviewee Ditlev Bredahl (founder of Hosted AI, his 7th startup) talks about why he can’t stop doing startups, and what true entrepreneurship takes out of you. He’s also a man with high EQ, unusual for a high achiever..

Let’s go shall we?

Vile Rumour & Gossip

• The race for an AEO feed is on… and it’s not fair. OpenAI has published the form for brands (or anyone) to complete so that GPT can find your products, but they’ve not opened it up yet. In the background ALL of the eCom platforms have done deals with the LLM’s to feed them - without clients being in the know.

• Beware the bill for your AI-created content - We have a few brands in our community who have adopted a pretty advanced landing pages strategy - creating 1000’s of landing pages using OpenAI. They are all being slightly screwed by the Search bill that’s coming their way - API calls maxed out and an escalated invoice. Much to unpick in the new world.

• Code quality of LLM’s (discussion points) - For all the SI’s who read this email, a very interesting trend is appearing in LLM’s writing code. Front end? Perfect - fine - excellent. But - the further you go down the stack the LLM’s are less effective, bloated even. Code isn’t fixed… it’s rewritten in order to bug fix. Customers are not interested in this.

Is Demand gen finally growing up?

To BDR or NOT to BDR, that’s the question. In the end everyone does, it’s too tempting and too baked in to funding models, but we’re seeing an explosion of newsletters, podcasts, owned conference series and EGC (employee generated content). The quality threshold is currently quite low, but like all things the hard yards will reap rewards… By issue 100 they’ll have some traction (they’ll all be binned by then right?)

Weirdly, at the same time as this is happening the quality of contact data in CRM’s across the SaaS market is declining as folks become reliant on the ABM machines like 6Sense. Highly dubious for GDPR btw but that ship has sailed now.

The last mile is the challenge still, that first meaningful conversation with a target account. We’re currently running a super-Bougie dinner series for several clients, with liberal amounts of restaurant temptation involved. We’re going back to the 80’s everyone!

The AI Solutions hunting a Problem

Of course it started with IBM Watson, great tech but no market fit. In the end clients couldn’t find a problem that it solved. Now we have 000’s of AI services launching into a market in which the customer is totally conflicted - let me explain.

The board sees “more with less”, cost savings and efficiency, but they don’t understand ANY of the tech. General staff are terrified of losing their role and so resist at every turn if possible. Engineering? Well they’re being told to automate, and internally are mostly a department under pressure for failing to deliver what the business wants 🙁 - so they’re under pressure to lose staff.

We are bound to see a whole ton of examples like Artifact, Ghost Autonomy and more who never get traction, but to the sales leaders out there tempted by the big VC cheque to go AI side… we’d be careful about anything before scaleup stage.

For clients - we’re hearing much more “wait and see”... unless the engineers are doing skunkworks to validate the tools and build trust internally.

eCom is fading

Is it just me, or is eCom on a sort of life support currently. While we wait for the LLM’s and google to decide how the transaction will change, physical encounters are increasing and stores in Europe are taking the investment once destined for digital.

Why? Well - when I speak to those operating retail brands, the revenue in store is currently more certain (and predicted to rise as people return to full days in an office), and tech projects typically do not fail, neither do they change mid-flight.

But the biggest factor right now is…. Tumbleweed : There’s real little advancement in eCom; attribution and tracking is all over the shop; the platforms aren’t innovating quickly in real terms, and so digital is being called out for being an unpredictable luxury channel in some brands.

Then of course there’s the specialists… BÆrskin, Swissklip, Snag, Pott’d and others are killing it because their focus is a super-tight coupling of product, creative and media choices… true test and learn with AI making everything faster and cheaper. But there are very very few of these.

Go where there is appetite… it’s tough to see much of it here.