{"id":18233,"date":"2026-08-18T11:41:42","date_gmt":"2026-08-18T09:41:42","guid":{"rendered":"https:\/\/incubeta.com\/?p=18233"},"modified":"2026-08-18T11:41:45","modified_gmt":"2026-08-18T09:41:45","slug":"what-is-the-moment-of-machine-mediation-momm-and-how-does-it-change-the-way-you-should-think-about-e-commerce","status":"publish","type":"post","link":"https:\/\/incubeta.com\/sg\/news-and-resources\/what-is-the-moment-of-machine-mediation-momm-and-how-does-it-change-the-way-you-should-think-about-e-commerce\/","title":{"rendered":"What Is the ‘Moment of Machine Mediation’ (MoMM), and How Does It Change the Way You Should Think About E-Commerce?"},"content":{"rendered":"\n

The “Moment of Machine Mediation,” or “MoMM,” is the point where discovery, recommendation, and purchase all collapse into a single, frictionless action. Paul Ruscoe<\/a>, VP of Marketing Intelligence at Incubeta takes a closer look, and shares his point of view.<\/em><\/strong><\/p>\n\n\n\n

Forgive the slightly clunky term, as if marketers needed another acronym to put into their vocabulary. But I do quite like it. “MoMM” could almost have a double meaning. Reflecting the childhood experience where your \u201cMoMM\u201d would dictate what you had for dinner, when and where you would go on vacation or, worse still, where you got your hair cut, what clothes you should wear or who you should date. I would hope that, like me, you’re glad you don’t have to delegate those latter choices now.<\/p>\n\n\n\n

Today, we are at the early stages of the next evolution in the buying process. You ask your model. It answers. In some circumstances, it buys. No results page, no comparison grid, no 40 minutes of open tabs and an abandoned basket.<\/p>\n\n\n\n

Conversational commerce reduces wandering. The zero-click economy is the version we already, to some degree, inhabit, where the answer arrives and the click never does. Frictionless commerce is what we experience when the answer also carries your card details. That is \u201ccustomer journey compression\u201d, but it is worth being precise about what is actually being compressed.<\/p>\n\n\n\n

What is being compressed is the corridor between memory and purchase. Which makes the width of a brand’s memory and accessibility the things that matters most.<\/p>\n\n\n\n

Think about sending someone to the bar for you.<\/p>\n\n\n\n

Say “get me a Guinness” and you get a Guinness. Say “get me a beer” and you get whatever they picked, which may be the thing they drink themselves, or the most familiar name on the tap. You are not standing at the bar. You cannot see the pumps. You cannot point at a bottle you half recognize and change your mind at the last second, which is what you could have done if you had walked up there yourself.<\/p>\n\n\n\n

Your pal becomes the mediator.<\/p>\n\n\n\n

Everything about what lands in front of you was settled by whether you had a brand in your head before you opened your mouth, or by what was available and known to be the least offensive choice to your palate. The opportunity to consider another brand disappears. Because you don’t care enough, certainly not consciously. And so familiarity and availability dictates choice.<\/p>\n\n\n\n

That is the human equivalent of the MoMM. That example, the pal in the bar, is also the limit of it.<\/p>\n\n\n\n

The person at the bar is carrying your drink, and they are not choosing your next home, your next car, or negotiating your next multi-year enterprise software contract.<\/p>\n\n\n\n

It is not collapsing the middle of the funnel<\/strong><\/p>\n\n\n\n

You will read a great deal about the messy middle imploding, the consideration phase evaporating, the funnel finally coming apart.<\/p>\n\n\n\n

Buyers do not assemble a preference by moving through stages of increasing engagement. They arrive already holding one, two maybe three primary choices. Jenni Romaniuk put this plainly for the B2B Institute in 2022, writing about Google rather than ChatGPT: memory-generated brands are the starting point for the buying process. Other sources, search engines, colleagues, are consulted only when the options that memory produced are insufficient. And when buyers do consult, they still lean toward the brands they already know. In fact, recommendation often reinforces choice, rather than redirecting it.<\/p>\n\n\n\n

The evidence is not thin. WARC reports research finding that consumers pick the brand they expected to buy in 84% of purchase journeys. In B2B, Harvard Business Review puts 80% of buyers as holding a day-one shortlist of vendors before they conduct any research at all, and 90% of those who buy end up choosing from that day-one list. Gartner adds that 40% of B2B buying processes end in no decision whatsoever. Only around 6% of buyers end up purchasing from a brand they discovered somewhere along the way.<\/p>\n\n\n\n

So the consideration set was never built in the middle of anything. It was built months or years earlier, by memory. Preference is a reflection of how easily your brand came to mind when the buying situation arose.<\/p>\n\n\n\n

Which means the MoMM is not eliminating a stage where minds were changed. It is eliminating a stage where minds were confirmed, and where brands that had already lost very occasionally got a second look. That second look was always a long shot. It could now become a longer one.<\/p>\n\n\n\n

It is mediation, not execution, and the scale is modest<\/strong><\/p>\n\n\n\n

Before anyone repositions their whole business around the MoMM, the numbers.<\/p>\n\n\n\n

McKinsey put up to $1tn of US B2C retail revenue as orchestrated by agentic commerce by 2030, with global projections of $3tn to $5tn. Treat that with the suspicion any consultancy forecast deserves. It is useful here only because it is close to the most bullish number in circulation, and James Hurman has since converted that to a share of retail. It comes out at somewhere around 10 to 14%. Even the optimists are therefore leaving 86 to 90% of your sales arriving the way they always have, through a human who already had you in mind.<\/p>\n\n\n\n

Add to this, how slowly buying behaviour actually shifts. E-commerce reached 16.9% of total US retail in the first quarter of 2026, over thirty years after Amazon opened for business. That is the share of retail happening online at all, never mind happening through an agent. Les Binet has made the same point, that after three decades of the internet, most purchases still take place offline and most involve no online research whatsoever.<\/p>\n\n\n\n

And the key word in the name is mediation. How much of the act the machine performs will vary with how much the category demands of the buyer, which is a question to answer for your own market before you do anything else. At the trivial end, replenishment and commodities, it will maybe execute. In the middle, cars, holidays, insurance, it will likely shortlist and a human buys, largely based on preference and price. At the complex end, anything with a committee, it does the reading and a room full of people still signs the contract.<\/p>\n\n\n\n

Hurman cites survey data from Salsify, who sell product data syndication and therefore have every reason to talk this up, suggesting that around a fifth of people are enthusiastic about agentic shopping. Another 40% are interested for the repeat grocery order and the utility bill, and explicitly not for electronics or clothes. The remaining 40% however, say shopping is personal and they are not handing it over. The enthusiasm is concentrated exactly where people would rather delegate the thinking. It\u2019s certainly not wholesale..<\/p>\n\n\n\n

People will not hand over the decisions they care about. They will not let a machine pick their car, their house, or the software their team has to live with for three years. They will not hand over the ones they enjoy either. Nobody is delegating the wine (if that is indeed, what they like!).<\/p>\n\n\n\n

What they will hand over is everything else. The dull, the repeated, the ones they stopped wanting to think about years ago. That is where the machine takes hold, and that is where its influence is greatest.<\/p>\n\n\n\n

None of what follows depends on any of these numbers. If agentic commerce arrives at three times the pace anyone is forecasting, the prompt still comes out of memory, and everything below will still likely hold.<\/p>\n\n\n\n

What it does change: the importance of your Category Entry Points<\/strong><\/p>\n\n\n\n

Somebody still has to type the prompt. What they type comes out of memory, a desire or a need state. “What is the best waterproof jacket that packs down small for hiking.”<\/p>\n\n\n\n

Three things are occurring. We have an instruction (What is the best\u2026), followed by a sub-category (waterproof jacket) linked to a need and buying situation (\u2026that packs down small for hiking).<\/p>\n\n\n\n

The machine then does what the brain would. It tries to filter information based on which brands match the buying situation.<\/p>\n\n\n\n

If your brand carries no association with the buying situation in the buyer’s head, you will, most likely, lose.<\/p>\n\n\n\n

If your brand carries no association with that buying situation anywhere in the body of scrapable text, the model has nothing to match against. And, you guessed it, you lose.<\/p>\n\n\n\n

That second gate is downstream of the first. Reviews, expert write-ups, retailer copy, forum threads, press coverage: all of that accumulates because people buy you and talk about you, which is a function of penetration. But that doesn’t mean you can relieve yourself from still having to feed the machine clean, accurate, structured product data. Because if the model cannot parse you, it cannot mediate on your behalf.<\/p>\n\n\n\n

So do it, and do it properly. Feeds, schema, product data, the lot. This is physical availability arriving in a new channel. It belongs to Easy to Find, and it is subject to the same rule that has always applied there. It removes friction for buyers who already had you in mind. Being Quick to Mind remains the other job, and it is the one that never finishes.<\/p>\n\n\n\n

The mechanism, with numbers on it<\/strong><\/p>\n\n\n\n

Jenni Romaniuk’s work on Category Entry Points measures how many buying situations buyers actually link to a brand. In US business insurance, State Farm, at 25% penetration, has 40% of decision-makers linking six or more buying situations to the brand, and 25% linking none. Hanover, at 5% penetration, has 8% linking six or more, and 64% linking no buying situations or need states with the brand at all.<\/p>\n\n\n\n

Sixty-four per cent of that market cannot therefore construct a prompt containing a buying situation that would return Hanover as the recommendation. Not from dislike. Hanover simply is not attached to anything they are trying to do.<\/p>\n\n\n\n

That is B2B, so before anyone waves it away as a quirk of insurance buyers, the same pattern turns up in the most consumer of consumer categories. John Dawes, also at Ehrenberg-Bass, reports toothpaste data where what he calls a brand’s equity score, meaning the proportion of the population who link any positive attribute to it, correlates with brand size. Colgate, at 37% penetration, scores 42%. Ultrabrite, at 7%, scores 21%. Yet among each brand’s own users the scores are near identical, hovering around 56% whichever brand you look at. The people who buy small brands feel much the same about them as the people who buy big ones. The difference between the two is simply how many people hold any association at all.<\/p>\n\n\n\n

Romaniuk and Sharp found the same across 28 categories and 208 brands. No small brand compensates with depth. There is no obscure challenger sitting on an unusually rich seam of associations, waiting to be discovered. Small brands are known by fewer people, in fewer situations. As Dawes puts it, mental availability is built in bits, across masses of people.<\/p>\n\n\n\n

So this is not an algorithmic phenomenon at all. It is the oldest finding in marketing science.<\/p>\n\n\n\n

Double Jeopardy picks up what I would suggest is a third punishment. Fewer buyers, slightly lower loyalty, and now surfaced less often by the machine. Algorithmic Jeopardy, if you like. The machinery underneath it is the width of your category entry point network. The algorithm is simply being obedient to the entry points associated with the brand.<\/p>\n\n\n\n

That is a hypothesis, and I want to be clear that it is not a finding. Nobody has yet run a thousand prompts across twenty categories and regressed citation frequency against penetration, which is precisely the study that would confirm it or kill it. Until someone does, I am extending a law rather than reporting one. What follows is what I think the evidence points to, offered in the hope that somebody with a research budget takes it away and proves me wrong.<\/p>\n\n\n\n

The corridor was a chance, and it was never a good one<\/strong><\/p>\n\n\n\n

In a browsing world, that 64% who had not associated Hanover with an entry point was still reachable. The results page gave Hanover a chance to be stumbled into. Someone with no association between Hanover and anything they were trying to do could still see the name, click the name, and buy the name.<\/p>\n\n\n\n

As we enter the MoMM, the stumbling stops.<\/p>\n\n\n\n

But being reachable is not the same as winning, and it is worth being honest about how much that chance was ever worth.<\/p>\n\n\n\n

The internet was supposed to obviate brand. Infinite shelf, then aggregators, then ratings and reviews from people who had actually bought the thing. Scott Galloway has made the point with hotels for years, that TripAdvisor let him stop defaulting to the big chains and find the perfect small hotel for his tastes every time. If that held at population scale, the share of the largest brands should have fallen as buyers dispersed toward perfect-fit obscurity.<\/p>\n\n\n\n

But then, as Galloway’s own numbers show, between 2000 and 2025, Marriott and Hilton more than doubled their market share and roughly tripled their profits.<\/p>\n\n\n\n

Twenty-five years of maximum findability, maximum comparison and maximum review data, and share still concentrated on the brands people already knew. Exposure was available. It was not sufficient. The brands that won the browsing era won it on the same thing they always won on, which is how many people held an association with them before the search box opened.<\/p>\n\n\n\n

Which is why the CEP network, as part of building durable memory, is a critical thing to fix. Findability was the weaker input even when the buying corridor was long. When that corridor truncates, the input that was already doing most of the work is about to do nearly all of it.<\/p>\n\n\n\n

The ceiling on the obvious workaround<\/strong><\/p>\n\n\n\n

There is a corollary, and it is a possibility rather than a probability. Prompts are long and specific, and specificity cuts both ways. A brand can be famous in general and completely absent against a particular question, which means a small brand with a genuine, earned advantage on one narrow and valuable buying situation could be surfaced ahead of brands much larger than itself when that question gets asked.<\/p>\n\n\n\n

Worth building. Worth measuring.<\/p>\n\n\n\n

But temper your expectations, because owning a buying situation outright is rare. Romaniuk and Gaillard, across eight categories and 130 brands, found that on average only 2% of consumers attach an attribute to one brand and one brand alone. Separate work across 17 categories found only around one in ten people regard any brand as different or unique at all. Distinctive positions in the mind are not lying around waiting to be claimed.<\/p>\n\n\n\n

And so the hard ceiling remains. You might own the narrow prompt. You will not own the broad ones, and the broad ones are where the volume lives. The bigger brand is likely to win, and win thrice. More market share, higher repeat rate, surfaced more often in the algorithm.<\/p>\n\n\n\n

For e-commerce specifically<\/strong><\/p>\n\n\n\n

Twenty years of the discipline have been built on intercepting people mid-wander. Paid search, shopping ads, retail media, product page optimisation, price comparison, retargeting, basket abandonment. Almost none of it creates demand. At best it stands in the corridor and catches demand as it walks past. At worst it stands next to demand that was already coming your way and takes credit for stopping it wandering off.<\/p>\n\n\n\n

The corridor will get shorter. Slowly, unevenly, possibly over a decade or more depending on the category, the market and the buyer, but in one direction. Every dollar aimed at being findable during the wander is aimed at a window that is narrowing, and the small to mid-market brands who have built their entire acquisition model on that window will feel it first, because they have the least memory to fall back on.<\/p>\n\n\n\n

The prompt is written from memory. Being linked to the buying situation before the buyer has one in mind is the input the machine cannot route around, and it remains the one nobody wants to fund.<\/p>\n","protected":false},"excerpt":{"rendered":"

The “Moment of Machine Mediation,” or “MoMM,” is the point where discovery, recommendation, and purchase all collapse into a single, frictionless action. Paul Ruscoe, VP of Marketing Intelligence at Incubeta takes a closer look, and shares his point of view. Forgive the slightly clunky term, as if marketers needed another acronym to put into their […]<\/p>\n","protected":false},"author":6,"featured_media":18311,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[217],"tags":[221,314],"approach":[],"solution":[],"industry":[50],"market":[19],"class_list":["post-18233","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-thought-leadership","tag-ai","tag-strategy","industry-global","market-us"],"_links":{"self":[{"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/posts\/18233","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/comments?post=18233"}],"version-history":[{"count":4,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/posts\/18233\/revisions"}],"predecessor-version":[{"id":18309,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/posts\/18233\/revisions\/18309"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/media\/18311"}],"wp:attachment":[{"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/media?parent=18233"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/categories?post=18233"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/tags?post=18233"},{"taxonomy":"approach","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/approach?post=18233"},{"taxonomy":"solution","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/solution?post=18233"},{"taxonomy":"industry","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/industry?post=18233"},{"taxonomy":"market","embeddable":true,"href":"https:\/\/incubeta.com\/sg\/wp-json\/wp\/v2\/market?post=18233"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}