DEEP BACKGROUND 11 OF 13 . THE PORTFOLIO Does tailoring the message to different groups actually work? Yes, and by less than the industry assumes. Across 57 studies and 58,454 people, tailoring produced r = .074. Against one generic message rather than against nothing, the advantage is about half what the headline suggests. WHY THIS IS WORTH MONEY Every established brand eventually faces the same choice. One message, made unmissable and repeated everywhere. Or one idea, expressed differently for people who relate to it differently. The first is cheaper to make and easier to buy. The second costs more and is usually argued for on taste. So it is worth knowing what the evidence says, including the parts that do not help us. WHAT THE RECORD SHOWS Tailoring works, and the effect is small Noar, Benac and Harris (Psychological Bulletin, 2007) meta-analysed the tailored print health communication literature: 57 studies, 58,454 people. "The sample size-weighted mean effect size of the effects of tailoring on health behavior change was found to be r = .074." That is a small effect, and we would rather print it than hide behind the word significant. Eight moderators mattered, including the type of comparison, the behaviour, the population and how many things were tailored on. And here is the comparison that actually matters to us Most tailoring studies compare tailored material against nothing. The question on this page is tailored against one good generic message, which is a harder test. Hao, Goetze, Alessa and Hawley (Journal of Medical Internet Research, 2023) separate the two. Tailored against general information: Hedges g = 0.16. Tailored against no information: g = 0.29. So expressing an idea several ways beats expressing it once, and by roughly half the margin that tailoring beats silence. Small, real, and worth exactly what it costs rather than more. Two larger meta-analyses agree on the size. Krebs, Prochaska and Rossi (Preventive Medicine, 2010), 88 studies and 106,243 participants: g = 0.17, declining over time but still significant at long follow-up, g = 0.12. Lustria and colleagues (Journal of Health Communication, 2013), 39 studies: d = 0.139. The mechanism argument, and its honest status The reason we hold that the two ends of a spectrum need different expressions is not the tailoring literature. It is that they appear to run on different transmission mechanisms. At the committed end, adoption behaves as a complex contagion, needing reinforcement from several independent sources, and it spreads through dense connection rather than long bridges. See pillar nine, where that is evidenced. At the wide end, what happens is closer to mere exposure. Repetition alone produces liking, with no argument and no new information. See pillar eight. Those are two different processes, not two degrees of one. A message optimised for one end is not merely weaker at the other. It is operating on a mechanism that is not running there. And the honest status of that argument: nobody has tested it. We could find no study that puts complex contagion and mere exposure side by side as competing mechanisms at two ends of one audience. It is an argument built from two well evidenced literatures, and it is presented as an argument. THE STRONGEST OBJECTION, AND THE ANSWER TO IT The Ehrenberg-Bass position, and this time it is aimed straight at the portfolio. Romaniuk, Sharp and Ehrenberg (Australasian Marketing Journal, 2007) examined perceived brand differentiation across seventeen categories in two countries. Eleven per cent of a brand's own buyers perceived it as different. Ten per cent as unique. Seventeen per cent as either. In their words: "most buyers of a brand do not see it as different or unique. Yet, these buyers still buy it." Their conclusion is to "place distinctiveness at the centre of brand strategy, where a brand builds unique associations that simply make it more easily identifiable." One consistent set of assets, used the same way, everywhere. That is the published root of the argument against varying anything. And there is a commercial field experiment that goes further. Lambrecht and Tucker (Journal of Marketing Research, 2013) ran an online travel firm's retargeting across 77,937 consumers over twenty one days, comparing a generic brand ad against a dynamic ad showing the specific hotel the person had browsed. The personalised version performed worse. "Exposure to regular generic retargeted ad doubles the probability to purchase for that day, but adding personalized content to this ad reduces the purchase probability by 67%." The answer has three parts and the first one is a concession. One. Distinctiveness and expression are different layers, and Ehrenberg-Bass is right about the first. The assets that make you identifiable should not vary. If a portfolio approach is used as licence to make four different-looking brands, it is being used wrongly and it will cost penetration. Two. What varies is what the idea is for, not what the brand looks like. Features, model, pricing and delivery vary. The truth underneath does not, and neither do the assets carrying it. That is one proposition with four entrances, not four campaigns. Three. Lambrecht and Tucker is the sharpest warning in this whole layer, and it is about timing rather than tailoring as such. Their own finding is that personalisation stopped underperforming once people's browsing showed their preferences had settled. Tailoring to a preference somebody has not formed yet reads as presumption. That is a real failure mode and it belongs on the page. WHERE THE EVIDENCE STOPS The tailoring meta-analyses are health behaviour change. Transfer to commercial marketing is an argument, not a measurement, and the only large commercial field experiment we could reach points the other way. The Noar moderator-level effect sizes were not obtainable. Only the headline r = .074 appears here, and no per-moderator figure should ever be quoted from us. The two-mechanism argument is not tested. Block 3 says so. It is the most original claim in this layer and it is currently reasoning rather than evidence. Nothing is attributed to How Brands Grow, which we did not read. The differentiation figures come from the peer reviewed paper. WHERE THE CONE SITS The question is not which position to pick. It is how the same idea shows up differently at each one, considering how features, model, pricing and delivery vary. The commonest mistake with an established thing is being aimed at the wrong end: over-serving the devoted, who would have stayed anyway, or overspending on the disengaged, who do not care. And the rule that governs it: it is the idea that moves through the cone, not the people, moved up it. You are not promoting anybody. You are building an idea good enough, and expressed well enough, to travel the whole spectrum on its own merits. WHAT TO DO ON MONDAY Take your current lead idea and write it four times, one per position, changing only how it is expressed, priced and delivered. If you cannot keep the same truth underneath all four, it is four ideas and you have a budget problem. Next is the engine underneath all of it, and why the room where the ideas are supposed to happen is the thing eating them. SOURCES Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133(4), 673-693. doi:10.1037/0033-2909.133.4.673 Hao, L., Goetze, S., Alessa, T., & Hawley, M. S. (2023). Effectiveness of computer-tailored health communication in increasing physical activity in people with or at risk of long-term conditions. Journal of Medical Internet Research, 25(1), e46622. doi:10.2196/46622 Krebs, P., Prochaska, J. O., & Rossi, J. S. (2010). A meta-analysis of computer-tailored interventions for health behavior change. Preventive Medicine, 51(3-4), 214-221. doi:10.1016/j.ypmed.2010.06.004 Lustria, M. L., Noar, S. M., Cortese, J., Van Stee, S. K., Glueckauf, R. L., & Lee, J. (2013). A meta-analysis of web-delivered tailored health behavior change interventions. Journal of Health Communication, 18(9), 1039-1069. doi:10.1080/10810730.2013.768727 Romaniuk, J., Sharp, B., & Ehrenberg, A. (2007). Evidence concerning the importance of perceived brand differentiation. Australasian Marketing Journal, 15(2), 42-54. doi:10.1016/S1441-3582(07)70042-3 Lambrecht, A., & Tucker, C. (2013). When does retargeting work? Information specificity in online advertising. Journal of Marketing Research, 50(5), 561-576. doi:10.1509/jmr.11.0503 Superfandom . Deep Background . https://superfandom.ai/science/the-portfolio/ Matt Hart . published 2026-09-08 . last changed 2026-09-09