AIDA

Origin

The usual story is that Lewis, an American advertising man, set out three steps in 1898 and added a fourth later. It comes from Edward K. Strong’s The Psychology of Selling and Advertising (McGraw-Hill, 1925), which says that in 1898 Lewis “used the slogan, ‘Attract attention, maintain interest, create desire,’” in an advertising course in Philadelphia, took the idea from William James, and “later on” added “get action” [1, p. 9]. Strong repeats the claim on page 348. He gives no source either time.

Strong was solving a training problem. America had become a buyer’s market and selling had to be taught: his count of salesmanship books runs from one in 1869 to 267 in the 1910s [1, p. 8]. AIDA was a mnemonic for salesmen and copywriters, never a finding about buyers.

The attribution has now been checked against Lewis’s own writing. Akinori Iwamoto read every surviving Lewis article and book from 1897 to 1910, including the Inland Printer columns written as Musgrove, and found no text in which Lewis prints the slogan [3]. The nearest is 1899: an advertisement “will attract, interest and convince” [3, Table 1]. The only record of Lewis saying the four words is a 1910 report of a speech in which he describes hearing “a writer” say them [3, p. 13].

The sequence was in print by then, under other names. Frank Dukesmith’s magazine Salesmanship set out “securing attention,” “arousing the interest,” “creating desire,” and “closing the sale” in June 1903 [3, Table 2]. Arthur Sheldon’s correspondence school taught attention, interest, desire and “resolve to buy” from 1903 and, by 1910, attention, interest, desire, action [3, Tables 2 and 5]. The first orderly printing of all four AIDA words Iwamoto could find is Seth Brown, December 1904 [3, p. 3]. Barry and Howard’s 1990 review, the standard academic history, repeats Strong’s version [5, p. 123, as quoted in 3], and Kotler’s Marketing Management has carried the Strong citation since 1972 [3, p. 2].

Established: the acronym and the Lewis story entered the literature through Strong in 1925; the four steps were a salesmanship convention by 1904; Lewis wrote in threes. Not established: that Lewis originated anything in 1898.

What the evidence shows

There is no empirical literature on AIDA as such. The tests exist for its descendant, the hierarchy of effects, so that is what can be reported.

Lavidge and Steiner (1961) turned the slogan into a measurement proposal. Their four-page Journal of Marketing article lists six steps, awareness, knowledge, liking, preference, conviction and purchase, groups them into cognitive, affective and conative functions, and argues for measuring advertising at each step rather than by sales alone [6][11, p. 8]. It is a theorist’s claim, not a finding: the article reports no test of the sequence. Palda (1966) coined “hierarchy of effects” and was the first to evaluate it [11, p. 2].

The largest review is Vakratsas and Ambler (1999). They read more than 250 journal articles and books and concluded that “there is little support for any hierarchy, in the sense of temporal sequence, of effects” [7, abstract]. They proposed dropping the sequence and studying advertising in a space of affect, cognition and experience, with position set by context [7].

The most direct test came twenty years later. Pauwels, Valenti, Srinivasan, Yildirim and Vanhuele modelled weekly data for 178 fast-moving consumer goods brands in 18 categories over seven years, with survey measures of cognition, affect and experience alongside sales [11]. Three results matter. A sequence does exist: for 94 per cent of brands an ordered model beat the simultaneous one, which fitted none of the 178 [11, p. 21]. The order is not AIDA’s: among the 90 brands where one sequence was statistically best, affect came first in 42 per cent, and the think-feel-do order AIDA assumes fitted 6 brands, 7 per cent [11, Table 6]. And the last factor in whatever sequence operates explains 56 per cent of the sales variation attributable to the three, the first only 16 per cent [11, p. 25].

Feel, think, use (ACE)38

Think, use, feel (CEA)17

Feel, use, think (AEC)15

Use, think, feel (ECA)6

Think, feel, use (CAE, the AIDA order)6

Use, feel, think (EAC)4

The order AIDA assumes, think then feel then act, was the best-fitting sequence for 6 of the 90 brands where any single order won.

Source: Pauwels, Valenti, Srinivasan, Yildirim and Vanhuele, 2020, MSI Working Paper 20-139, Table 6, n = 178 brands, 18 categories, 7 years. Redrawn by Brave New Media.

Unit: brands

Show the data

Item	brands
Feel, think, use (ACE)	38
Think, use, feel (CEA)	17
Feel, use, think (AEC)	15
Use, think, feel (ECA)	6
Think, feel, use (CAE, the AIDA order)	6
Use, feel, think (EAC)	4

Ehrenberg reached a similar shape by another route in 1974. His awareness, trial, reinforcement account holds that for established brands of frequently bought goods advertising’s “main role is to reinforce feelings of satisfaction with brands already bought” [12]. The Ehrenberg-Bass Institute’s position today, in Sharp’s words: “The dominant way that advertising works is by refreshing, and occasionally building, memory structures that improve the chance of the brand being recalled and/or noticed in buying situations,” and “most sales typically come from people who had not formed an intention” [13]. On this account there is no desire stage. There is memory, and there is the buying occasion.

Industry survey data describe the same buyer. In the 2024 Edelman-LinkedIn study of 3,500 B2B decision makers, 75 per cent said one piece of thought leadership had led them to research a product or service they were not previously considering, and 70 per cent of C-suite respondents said one had made them question whether to keep a current supplier [19]. The content did not find an interest stage. It made one.

Researched a product they were not previously considering75%

Of those: realised they were missing a significant opportunity60%

C-suite: questioned continuing with a current supplier70%

Of those: ended or significantly reduced a supplier relationship25%

Of those who researched: began buying from the publisher23%

The buyers who moved had formed no intention to move: the content made the interest stage rather than finding it, which is the Ehrenberg-Bass reading, not AIDA's.

Source: Edelman and LinkedIn, 2024 B2B Thought Leadership Impact Report, 2024, n = 3,500 B2B decision makers and C-suite executives in seven countries, key findings as published by Edelman, 28 Feb 2024. Redrawn by Brave New Media.

Unit: per cent of respondents

Show the data

Item	per cent of respondents
Researched a product they were not previously considering	75
Of those: realised they were missing a significant opportunity	60
C-suite: questioned continuing with a current supplier	70
Of those: ended or significantly reduced a supplier relationship	25
Of those who researched: began buying from the publisher	23

Binet and Field’s analysis of 996 IPA Effectiveness Awards cases separates the two mechanisms AIDA fuses [14, p. 10]. Rational messaging produces the bigger short-term sales effect and decays; emotional priming produces a smaller immediate effect that accumulates [15, slides 13 and 14]. Emotional campaigns reporting very large profit growth rise from 13 per cent at one year to 43 per cent at three years or more; rational campaigns from 10 to 23 per cent [14, Figure 52][15, slide 16]. “Feelings towards brands outlast memories of facts and figures about the brand” [14, p. 53].

Cialdini addresses the first letter. Pre-Suasion argues that the factor most likely to determine a choice “is the one that has been elevated in attention (and thereby in privilege) at the moment of decision” [16, chapter 2 summary]. To get action “it’s not necessary to alter a person’s beliefs or attitudes or experiences,” only “what’s prominent in that person’s mind at the moment of decision” [16, chapter 1]. The principles of Influence are likewise levers, not stages [17]. Attention, for Cialdini, is not the entrance to a corridor. It is the decision.

The challenge

The empirical challenge is real but aimed at the family, not the member. Nobody has published a test of AIDA because there is nothing in it to test: four states and an asserted order, with no measures and no timing.

The strongest published critique is Weilbacher’s 2001 Journal of Advertising Research essay: hierarchy models do not describe what advertising does to people, do not describe how it works in markets, and while advertising is thought about in hierarchy terms its true effects will not be measured [8]. Barry replied in defence of the hierarchy as a guide for practice and research [9]. Weilbacher’s rejoinder: “If a widely accepted theory has not been proven over the course of 100 years, isn’t it time to say so?” [10, p. 49, as quoted in 11, p. 3].

The positions are worth stating exactly. Vakratsas and Ambler: no temporal hierarchy is supported [7]. Pauwels and colleagues: a hierarchy is supported, it varies by brand and category, affect usually leads, and the fully simultaneous model fits no brand [11]. Ehrenberg and Sharp: the intermediate states are mostly beside the point, because advertising works on memory and most buying happens without prior intention [12][13]. Binet and Field: emotional and rational effects run on different clocks, so any single sequence is a category error [14]. Barry: the hierarchy remains useful as a planning device [9].

Two things are not in dispute. No body of evidence supports a fixed attention-interest-desire-action order across brands. And every camp agrees that purchase feeds back into the first step, which a one-way sequence cannot represent [11, p. 2][12].

Where practitioners get it wrong

Reading it as a journey. AIDA describes what one message should do. The funnel reads it as what one buyer does, over weeks, across channels, one stage at a time. The funnel is a practitioner convention with no author and no test. Stage-gated content, where a prospect is “moved” from awareness assets to consideration assets before seeing an offer, is built on it. Pauwels’ data say the first factor explains 16 per cent of sales variation and the last 56 per cent [11, p. 25]; Sharp’s say most buyers never formed an intention [13]. Gartner’s survey of 750 B2B buyers found that buying groups spend 17 per cent of a purchase meeting any potential supplier and 45 per cent researching on their own, that a complex purchase involves six to ten decision makers, and that the six buying jobs are revisited in loops rather than completed in order [18].

Researching independently online27%

Meeting with the buying group22%

Researching independently offline18%

Meeting with potential suppliers17%

Other16%

The supplier is present for a sixth of the purchase and any one rep for 5 or 6 per cent of it; the corridor AIDA describes is walked, if at all, out of the marketer's sight.

Source: Gartner, B2B Buying Journey, 2019, n = 750 B2B buyers, infographic CM_611049. Redrawn by Brave New Media.

Unit: per cent of buying group's time

Show the data

Item	per cent of buying group's time
Researching independently online	27
Meeting with the buying group	22
Researching independently offline	18
Meeting with potential suppliers	17
Other	16

The channel side of the picture is the same. McKinsey’s B2B Pulse counted the distinct channels a buyer uses across one purchase at five in 2016, 7.5 in 2019 and ten in December 2021, and in its April 2024 wave of 3,942 decision makers found between 30 and 37 per cent of time in each of in-person, remote and digital self-service at every one of four buying stages, whichever buyer type was asked [20]. No stage owns a channel, so a content plan that assigns channels to stages is assigning them to a fiction.

20165

20197.5

Dec 202110

A buyer who meets the brand in ten channels, a third of the time in each mode at every stage, is not moving down a corridor with one channel per step.

Source: McKinsey, Global B2B Pulse, The new B2B growth equation, 2022, Exhibit 4, n = 1,042 (2016), 1,420 (2019), 3,360 (Dec 2021). Redrawn by Brave New Media.

Unit: channels

Show the data

Item	channels
2016	5
2019	7.5
Dec 2021	10

Reading it as a measurement plan. Dashboards that report awareness, consideration, intent and purchase as a cascade assume the order they are meant to test. Attitude measures are unstable: the Ehrenberg-Bass Institute reports about a 50 per cent chance that a person states the same attitude to a brand in two surveys [13]. The leaks in such a cascade are partly measurement noise.

Reading attention as a hook for an argument. The AIDA copy formula treats attention as the door and the rational case as the room. Binet and Field’s rational campaigns win at six months and lose at three years [14, p. 53]. Cialdini’s work says what is in mind at the moment of decision does the deciding [16]. The error is spending the creative budget on the argument and the media budget on the hook. HubSpot’s survey of 1,400 sales professionals reports that 96 per cent of prospects have researched before their first conversation with a rep, and 88 per cent already know the company [22]; the hook has been and gone before the seller arrives.

Did their own research before speaking to a rep96%

Know about the company at the first conversation88%

Know how the products compete with others84%

Know about the products or services83%

Prefer their own research to speaking to a human71%

Attention and interest have happened before the seller is in the room; the message that meets the buyer there is not the one that hooked them.

Source: HubSpot, 2024 Sales Trends Report, 2023, n = 1,400+ sales professionals in seven countries, pp. 10 and 15. Redrawn by Brave New Media.

Unit: per cent of prospects

Show the data

Item	per cent of prospects
Did their own research before speaking to a rep	96
Know about the company at the first conversation	88
Know how the products compete with others	84
Know about the products or services	83
Prefer their own research to speaking to a human	71

Our own observation. In the smallest of the supply chain group’s four markets, prospects doubted the group had a genuine local presence, and sales spent the first ten minutes of every call proving it did. We filmed their own people at their own site in the local language, no studio, no script. Engagement ran more than three times that of produced static work, and it cost less to make. The commercial effect sat elsewhere: prospects arrived already past the first step. The sequence did not run attention, interest, desire. Proof removed a question, and the conversation started further in.

What to do with it

1. Write AIDA into the brief as a copy checklist for a single asset, nowhere else. Check: the brief’s journey section cites something other than AIDA.
2. Remove any gate that withholds an offer until a prospect has consumed earlier-stage content. Check: every prospect can reach the offer from any first touch.
3. Measure cognition, affect and experience in the same quarterly survey wave and record which moves first for your brand. Check: all three sit in one instrument.
4. Set reach targets on the category, not the CRM. Check: the media plan states a share of category buyers reached, including non-customers.
5. Run two evaluation clocks: activation copy at three to six months, brand priming at three years. Check: each campaign line carries its window.
6. Before writing any message, write one line on what the audience will be attending to at the moment of exposure. Check: the line is in the creative brief.
7. Stop citing Lewis 1898. Cite Strong 1925 for the acronym and Dukesmith, Sheldon and Brown 1903 to 1904 for the sequence. Check: the deck footnote.