Consumer Shifts Are Systems, Not Trend Lists

BCG's latest consumer research highlights how structural change can reach across categories. GLP-1 provides a concrete case for why the relationships between effects can matter as much as the effects themselves.

The catalyst: BCG's consumer research points past categories

In August 2026, BCG surveyed more than 13,000 consumers across 12 markets and identified five durable shifts in how people live and buy.¹ One observation matters especially for the intelligence problem: these shifts do not sit neatly inside the categories companies use to organize research and strategy.

That raises a harder question than "what is trending?" What does an intelligence system need to represent when one change shows up across several parts of consumer life at once?

GLP-1 provides one worked example. It does not prove that every consumer shift behaves as a system. It does show how one shift can produce connected effects that are easy to flatten into separate category trends.

GLP-1 shows how one consumer shift can behave as a connected system rather than a list of isolated trends.

The limit of the trend list

Many intelligence systems are organized by category, team, or research question. GLP-1, viewed that way, generates a food trend, a wellness trend, an apparel trend, and a restaurant trend — each tracked by the team that owns it, and each observation possibly correct.

The problem is what separation hides: that one effect runs through another rather than alongside it, that two credible sources point in opposite directions, or that one connection rests on a product label while another rests on what respondents say they intend to do. Category research, surveys, and social listening all see real things. The trend list is not wrong so much as flat — and the connected view is a complement to that work, not a replacement for it.

The GLP-1 consumer system

The spine of the system is a mediated chain. For semaglutide, the Wegovy prescribing information describes regulation of appetite and reduced caloric intake, alongside body-weight reduction.² Changing body size is then associated, in observed purchase data, with apparel and wardrobe behavior.³⁴

GLP-1 use → appetite and food quantity → body weight and size change → apparel, fit, and wardrobe.

The interpretation matters more than the diagram. The apparel effect is not represented as a direct pharmaceutical effect; the commercially important relationship runs through the physical outcome of changing body size. A trend list records "GLP-1 affects apparel." The system records how.

Around that spine sit the other established relationships: GLP-1 use to grocery mix and spend, where an observed household spend decline and a reported mix shift are kept as two different measures rather than averaged⁴⁵; to protein demand, strictly as valuation⁶; to restaurant ordering, as stated behavior⁷; appetite and food quantity related to restaurant ordering composition, drawn without a causal arrow⁷; and to household routines, where users report behavior changes extending beyond the individual on the medication.⁵

The map below is a retrospective, descriptive worked example — eight nodes, eight relationships, assembled from named public sources. It is not a prediction, not a timeline, and not a claim that Datapiphany discovered any of these effects. They are public. The contribution is the structure.

Not all evidence is equal

Every connection on the map carries an evidence class, because the connections are not equally established. Class A is causal or mechanistic — here, the prescribing information. Class B is an observed behavioral association — purchase panels and matched demand analysis that show a relationship without proving the medication caused it. Class C is stated behavior — what consumers report doing or intending.

The map also combines regulatory, peer-reviewed, panel, consultancy, and industry-survey evidence. Those sources are useful, but they do not carry identical evidentiary weight — which is why mechanism, observed behavior, and reported behavior stay visually and verbally distinct. The practical rule for a reader is short: relationship is not causality. A map that draws every connection with the same line is making a claim it cannot support.

Three places the connected view changes the question

The commercial value of the structure is not a set of recommendations. It is a different, more decision-specific question in each category the system touches.

Food and nutrition: valuation is not volume. For semaglutide, reduced intake is already part of the established mechanism. The less obvious commercial question is what happens when valuation and volume diverge. The less obvious finding is that peer-reviewed demand analysis shows GLP-1 users exhibiting higher willingness-to-pay and more inelastic demand for most protein products.⁶ That is a valuation result, not a volume result: it says nothing about more pounds sold. Volume and valuation can diverge, and only a view that keeps them distinct can ask: where is valuation moving differently from volume — and where should we defend price rather than pounds?

Apparel: the mediated chain is the insight. Observed data supports the size-mediated path: a majority of active users in one commercial study purchased clothing or footwear primarily because their size changed, most anticipate broader wardrobe replacement, and a household panel shows apparel spend rising in the months after adoption.³⁴ Apparel research on its own can find some of this. What the joined view adds is that the same underlying shift also touches grocery baskets, restaurant ordering, protein valuation, and household routines — so the question becomes: how might changing body size alter size curves, replacement cadence, channels, fit, and shopping experience?

Restaurants: the contradiction is the finding. The restaurant evidence does not resolve, and the map does not pretend it does. In one industry survey, GLP-1 users reported more current restaurant occasions than non-users — even as nearly half said they were eating out less often than before starting medication, and most reported changing what they order: smaller portions, more protein, more vegetables.⁷ A separate household panel shows quick-service restaurant dollars declining among GLP-1 households.⁴ None of this supports "restaurant demand is rising," and none of it supports "restaurant demand is collapsing." A decision-intelligence system should preserve contradiction when the evidence does not support resolution. The usable question: how might ordering occasions, portion architecture, protein needs, and beverage choices change while restaurant participation persists?

What we did not draw

Several candidate relationships were investigated and are not on the public map. Alcohol: the strongest evidence is clinical, from trials in alcohol-use-disorder populations, and that population scope does not match a general consumer map. Fitness: the evidence did not establish a clear net direction. Beauty and aesthetics: available evidence did not meet the established-edge threshold. Spending reallocation: users state intentions to redirect part of reduced food spending, but stated intentions did not justify visually tracing actual dollar flows. The observed QSR spend decline, likewise, was retained as one dataset without being converted into a resolved restaurant-demand direction.⁴⁵

This is evidence discipline. Sometimes the most important analytical decision is refusing to draw the edge.

What this map demonstrates — and what it doesn't

It demonstrates that GLP-1 is associated with demand and behavior shifts across multiple consumer categories; that at least one commercially important relationship is mediated through another outcome rather than direct; and that fragmented public evidence can be organized into an inspectable relationship structure that retains evidence classes and contradictions.

It does not establish forecast accuracy, lead time, or the order in which effects arrived. It does not prove that every consumer shift behaves this way, that relationship structure beats category research, or that this approach is unique to Datapiphany. Those are different claims, requiring different proof.

From trend knowledge to decision structure

Datapiphany structures fragmented cultural evidence as relationships among signals, communities, behaviors, meanings, and categories, then evaluates those relationships against a specific growth decision. The working sequence is Signal → Relationship → Meaning → Opportunity → Decision.

The output that matters is not a longer trend list. It is a more useful decision frame. Not "GLP-1 is a food trend," but where are volume and valuation separating? Not "GLP-1 affects apparel," but how does size change alter fit, assortment, and replacement behavior? Not "GLP-1 threatens restaurants," but which parts of participation, ordering composition, and spend are actually changing?

The harder question

For a shift like GLP-1, the useful growth question is not only "what trends should we know about?" It is: what is connected, what does the connection mean, how certain are we, and what decision changes because we can see it? GLP-1 is one worked example of what an answer looks like when the evidence — including its conflicts and its refused edges — is allowed to keep its shape.

If you're trying to understand how an emerging consumer shift could affect your category, audience, or portfolio, Datapiphany can map the evidence against the decision you need to make.

Sources

Boston Consulting Group, Global Consumers Have Moved On. Has Your Growth Strategy Caught Up? (26 August 2026). Structural context; not a GLP-1 graph edge.
U.S. Food and Drug Administration / Novo Nordisk, Wegovy (semaglutide) U.S. Prescribing Information.
Circana, GLP-1 Medication Usage is Impacting the U.S. Apparel Industry (23 March 2026).
PwC, The end of more: How GLP-1 users shop, eat and seek support (22 June 2026).
Boston Consulting Group, The Unexpected Ways GLP-1s Are Transforming Consumer Behavior (6 August 2026).
Bina, Tonsor, and Richards, GLP-1 use and protein demand, Food Policy 138 (January 2026).
National Restaurant Association, GLP-1 and Restaurants: Shifting Habits, Not Shrinking Demand (20 May 2026).

Consumer Shifts Are Systems, Not Trend Lists

The catalyst: BCG's consumer research points past categories

Consumer Shifts Are Systems, Not Trend Lists

The limit of the trend list

The GLP-1 consumer system

The GLP-1 Consumer System Map

Not all evidence is equal

Three places the connected view changes the question

What we did not draw

What this map demonstrates — and what it doesn't

From trend knowledge to decision structure

The harder question

Sources