The Long and the Short of It is a 2013 report by Les Binet and Peter Field, published by the IPA (the UK’s Institute of Practitioners in Advertising), which analyzed nearly a thousand campaigns in the IPA Effectiveness Databank and concluded that long-term brand building and short-term sales activation are two different kinds of marketing that work over different timescales. Its best-known finding is that, on average, campaigns were most effective when roughly 60 percent of the budget went on brand building and 40 percent on activation, which became known as the 60:40 rule.
It’s hard to overstate how influential the report has been: I’ve seen its charts in more strategy decks than any academic paper on advertising, and “Binet and Field” has become shorthand for the case for brand. That popularity has a cost, though, because a lot of what gets quoted (a single ratio, a single chart) leaves out the context and the caveats the authors themselves were careful about. As someone who builds marketing mix models, I also think the report contains a direct challenge to how most models are built, which I’ll come to at the end. First, what the research actually did and found, including the later refinements and the share of voice rule of thumb that goes with it.
What Binet and Field studied in the IPA Databank
The IPA Effectiveness Databank is a collection of case studies submitted as entries to the IPA Effectiveness Awards, a competition that has run for over 40 years, in which agencies and advertisers document a campaign’s strategy, budget, media and business results. By 2019 it held around 1,500 cases, according to Binet and Field. For The Long and the Short of It, an update of their 2007 report Marketing in the Era of Accountability, they ran a meta-analysis across 996 campaigns for 700 brands in 83 categories and compared which strategies were associated with success over the short term and the long term.
Two measures run through all their work, and it helps to know what they are before reading any of the charts. The first is the number of “very large” business effects a case reports, across outcomes like sales, market share, profit, penetration, loyalty and price sensitivity, where only the top box of a four-point scale counts. The second is ESOV efficiency, the market share growth a campaign achieved per year for each 10 points of excess share of voice (more on ESOV below), which lets them compare campaigns with very different budgets. Most of the findings are differences in these measures between groups of cases, so they’re associations across the databank rather than measured effects of a single campaign.
Long-term brand building vs short-term sales activation
The central idea is that marketing has two kinds of effect. Sales activation (direct response, promotions, search, rational messages aimed at people ready to buy now) produces immediate, measurable sales spikes that decay quickly. Brand building (broad-reach, usually emotional advertising aimed at everyone who might buy in future) produces smaller immediate effects, but they accumulate over time into stronger brand preference, higher baseline sales and lower price sensitivity. You need both, Binet and Field argue, because there are no long-term effects without short-term ones, but the short-term effects are what make campaigns look efficient while the long-term effects are what drive growth and profit.
The report turns this into ten principles, which in my own words come down to the following. Effective marketing supports both sales volume and the ability to charge a premium, and the pricing benefits take years to show up. Campaigns should talk to everyone in the market, not just existing customers, because targeting the whole market was associated with far more very large effects than targeting existing customers. Emotional messaging builds the long-term effects while rational messaging drives the short-term response, and campaigns that aim for fame (getting people talking about the brand) were associated with about four times the ESOV efficiency of other campaigns (1.4 against 0.3). Creativity improves efficiency, share of voice matters more than ever, brand and activation work best when integrated (cases using both sets of channels showed roughly double the ESOV efficiency of brand channels alone), budgets should balance the two, and measurement should cover both time horizons rather than trusting short-term metrics alone.
The time dimension is the part that matters most for measurement. The report shows the share of cases reporting very large effects rising with campaign duration, particularly for profit and price effects and particularly for emotional and fame-driven campaigns, while rational campaigns did relatively well on short-term direct effects. The uncomfortable implication is that the campaigns that look best in a three-month evaluation are often not the ones that look best after three years.
Where the 60:40 rule comes from, and how it was refined
The 60:40 rule comes from a single chart in the 2013 work, plotting the number of very large effects against the share of budget spent on activation, with the optimum at roughly 40 percent activation. It has been restated and refined several times since, and the refinements are worth knowing because they’re where most of the nuance lives. When Media in Focus came out in 2017, updating the analysis for the digital era, Field summarized it as “around 60% long-term brand building and 40% short-term activation still the best combination” (B&T, 2017). Effectiveness in Context (IPA, 2018) then showed how the ideal balance shifts with the situation: across for-profit cases from 1998 to 2016 the curve peaked at 62 percent brand, with a 56 percent loss of effectiveness for an all-activation budget and a 20 percent loss for an all-brand one. The same report argued that when activation is easy (because customers research and buy online, because the business is subscription-based, or because the product is genuinely new) you can down-weight it, and vice versa. Finally, The 5 Principles of Growth in B2B Marketing (2019), written for LinkedIn’s B2B Institute, put the B2B optimum at around 46 percent brand and 54 percent activation, while stressing that the ratio is a guiding principle from a small sample rather than something to follow precisely.
These are the published optimum splits I’d use as starting points:
| Context | Brand : activation | Source |
|---|---|---|
| B2C average | 60 : 40 | The Long and the Short of It (2013) |
| All for-profit cases, 1998-2016 | Peak at 62 : 38 | Effectiveness in Context (2018) |
| Durables | 58 : 42 | Effectiveness in Context (2018) |
| FMCG | 60 : 40 | Effectiveness in Context (2018) |
| Financial services | 80 : 20 | Effectiveness in Context (2018) |
| Other services | 51 : 49 | Effectiveness in Context (2018) |
| Retail | 64 : 36 | Effectiveness in Context (2018) |
| B2B | 46 : 54 | The 5 Principles of Growth in B2B Marketing (2019) |
The shape of the 2018 curve is as important as its peak. It’s fairly flat near the top and falls away much faster on the activation side than on the brand side, so being 10 points off the optimum costs little, while putting almost everything into activation costs a lot. That asymmetry, more than the exact ratio, is the practical message for a team that has drifted into spending nearly all its budget on performance channels.
Share of voice and excess share of voice
Share of voice (SOV) is your brand’s share of all the advertising in your category, usually measured as your ad spend divided by total category ad spend over the same period. The term gets used loosely for share of impressions, search visibility or social mentions too, and those can be useful proxies, but the rules of thumb below were built on advertising spend. Excess share of voice (ESOV) is your share of voice minus your share of market (SOM): a brand with 10 percent of category ad spend and 8 percent market share has an ESOV of plus 2 points.
The relationship between the two is one of the oldest observations in advertising. It’s usually traced to John Philip Jones’s 1990 Harvard Business Review article “Ad Spending: Maintaining Market Share”, which compared brands’ share of voice with their share of market, and Binet and Field describe the share of voice rule as having been known for about 50 years. The rule is simple: brands that set SOV above SOM tend to grow, brands that set it below tend to shrink, and the rate of growth is roughly proportional to ESOV. Binet and Field quantified it from the IPA Databank. For consumer brands, 10 points of ESOV was associated with market share rising by about 0.6 points a year, all else being equal, and for B2B brands the figure was 0.7. The variation between categories is bigger than the gap between B2C and B2B: about 0.2 points per 10 points of ESOV in retail, 0.3 in FMCG, 0.6 in durables, 1.5 in financial services and 1.8 in other services.
The calculator applies those efficiencies to your own numbers, and its projection shows something the simple rule hides: if you hold your spend flat, your share of voice stays the same while your market share rises towards it, so your ESOV and your growth rate shrink each year. Growing share for several years in a row means increasing spend as you grow. The second half of the widget checks your brand and activation split against the published benchmarks, using the shape of the 2018 curve; it’s a rough guide to how far you are from the average optimum, not a measure of your own brand’s response.
Criticisms and limits of the Binet and Field research
The biggest limitation is the data. Every case in the IPA Databank was entered into an awards competition, so the sample is biased towards campaigns their authors thought were successful, and Binet and Field say so themselves, along with the fact that the cases skew towards the UK and towards relatively big budgets (they argue that comparisons with other research suggest the bias is modest). The effects are reported by the case authors on a top-box scale, and the analysis compares groups of cases rather than running experiments, so the findings are correlations: brands that spent more on brand building also reported more very large effects, but brands that could afford to spend on brand building may differ in other ways too. Byron Sharp has been reported as calling the 60:40 rule “terrible and very misleading” and arguing that a dataset of awards submissions is flawed (Mumbrella, 2022).
The second limitation is definitional. “Brand” and “activation” are categories of objective and style as much as channel, and in practice teams sort spend by channel (TV is brand, search is activation) in ways that don’t always match how the money works. A YouTube campaign optimized for conversions and a search campaign on generic category terms don’t fit neatly into either box. The third is that a cross-brand average is not a forecast. A rule of thumb like 0.6 share points per 10 points of ESOV tells you what happened on average across award entries, not what will happen to your brand with your creative, distribution and pricing, and the authors’ own context work shows how much the optimum moves. I treat these numbers as priors (a reasonable starting expectation before your own evidence arrives) rather than targets, and the ratio as a prompt to check whether your mix has drifted too far towards activation, not as a formula.
What The Long and the Short of It means for marketing mix modeling
This is where the report challenges how models are built. A typical MMM is fitted on two or three years of weekly data, uses adstock decay rates measured in weeks, and attributes everything it can’t explain with media and controls to baseline sales. By construction, that setup is good at measuring the short-term effects Binet and Field call activation and poor at measuring the long-term effects of brand building, because those show up as a slow rise in the baseline and a gradual fall in price sensitivity, which the model treats as trend rather than as something marketing caused. The result is a systematic bias: short-window models undercount brand, overstate the relative ROI of activation channels, and if you follow their budget recommendations blindly you drift towards exactly the activation-heavy mix the IPA data warns against.
There are practical ways to reduce that bias, none of them perfect. You can allow longer carryover for brand channels and check whether the data supports it, model the baseline as something marketing can move (for instance with a brand-awareness or consideration series as an intermediate outcome, or a second-stage model of base sales over a longer history), track price elasticity over time as a long-term outcome (the 2013 presentation shows DDB Matrix data in which a brand with high ESOV saw its price elasticity weaken markedly over three years), and use long-running geo holdouts to measure brand effects directly. At the very least, report the short-term ROI as a lower bound for brand channels and say so explicitly when you hand the numbers to someone making a budget allocation decision. In my experience the cheapest improvement is simply being honest in the decision memo about what time horizon the model can see.
None of this means the model is wrong and the IPA is right, or the other way round. They measure different things over different horizons, and the most useful measurement framework uses each for what it’s good at: experiments and models for the short-term response, and longer-horizon evidence, including benchmarks like Binet and Field’s, for what brand building does over years. If you want to see where this report sits alongside the academic work on advertising measurement, such as Lewis and Rao on how noisy ad effects are, the MMM research papers page puts them side by side, and the Vexpower marketing mix modeling course works through base and incremental sales on its invented Northstar brand.
