Essay

Marketing in a recession: why cutting is a bet, and what four downturns show

Published · Updated 18 min read

Marketing in a recession is the set of decisions a business makes about its advertising and brand spending when demand falls, and above all whether to cut, hold or raise that spending while competitors are cutting. Cutting feels like saving money. In practice it is a bet: that rivals will go quiet too, and that any share lost can be won back cheaply when demand returns. The best evidence on that bet comes from two places, the Ehrenberg-Bass Institute's studies of brands that stopped advertising, and Peter Field's analysis of IPA effectiveness cases from the 2008 to 2009 recession.[1][2]

This essay brings those studies together with our own analysis of public US data. Using the National Bureau of Economic Research recession dates as published on FRED, we tracked advertising jobs, advertising industry revenue and advertising prices through the four US recessions since 1990, as far as each series reaches back. The result shows why a downturn is when share of voice gets cheaper, and why the firms that keep spending face less competition for attention. We close with a framework, the Cut-or-Hold Matrix, and a step-by-step way to cut waste without cutting presence.

Idea in brief

The problem. In a downturn, marketing is one of the few costs a business can switch off without firing anyone. So it goes first, and the cost shows up a year or two later.

Why it happens. The saving is visible this quarter. The loss is not. Brands that stopped advertising lost on average 10% of their market share after one year, 20% after two and 28% after three.[1]

The gap (our data). In three of the last four US recessions, advertising jobs fell 2 to 8 times as far as jobs overall. In 2009, advertising services revenue fell 13.7% year on year at its worst quarter; nominal GDP's worst quarter was a 3.3% fall. When everyone cuts, the brands that stay are louder for the same money, and in spring 2020 the price of US ad space and time fell 8.1% in four months.

The answer. Cut waste, not presence. Keep share of voice at or above share of market if cash allows, and use the cheaper attention to build share for the recovery.

Key findings

The essay in 3 minutes
  1. Finding 128%

    Going dark costs share, steadily.

    Across 365 brands in 22 US supermarket categories, brands that stopped advertising lost on average 10% of market share after one year, 20% after two and 28% after three. Small brands and brands already in decline lost more.[1]

  2. Finding 24.5%

    Brands that pushed share of voice in 2008 to 2009 grew faster.

    Among about 50 IPA cases from the recession, brands with extra share of voice above 8% grew market share by 4.5% a year, against 1.0% for brands at or below zero.[2]

  3. Finding 3167

    The same spend works harder in a recession.

    For the high-investment group, market share growth in the recession indexed at 167 against cases from the two years either side. For the under-investors it was 46.[2]

  4. Finding 415.1%

    Advertising activity falls harder than the economy.

    In the 1990 to 1991, 2001 and 2008 to 2009 recessions, US advertising and related services jobs fell 5.2%, 15.1% and 14.3%, against 1.4%, 2.0% and 6.3% for all jobs. (Our analysis of BLS data.)

  5. Finding 5−8.1%

    Attention gets cheaper.

    From February to June 2020, the US producer price index for advertising space and time fell 8.1%, against an average rise of 0.6% over the same months in 2015 to 2019. (Our analysis of BLS data via FRED.)

  6. Finding 681 months

    Recovery is slow.

    After the 2008 to 2009 recession, advertising jobs took 81 months to regain their pre-recession level. After 1990 to 1991 they took 53 months, against 31 for all jobs.

Chapter 1 · 1 min

What marketing in a recession involves

A recession changes the economics of attention. Demand falls, budgets get cut, and the total amount of advertising in many categories shrinks. That has two effects. A business that goes quiet starts to be forgotten. A business that keeps going gets a larger slice of a smaller conversation, often at a lower price.

Terms used in this essay

TermMeaning
RecessionA significant decline in economic activity, dated in the US by the National Bureau of Economic Research. FRED's USREC series marks the months.
Going darkStopping all advertising for a period, usually a year or more.
Share of voice (SOV)A brand's share of all advertising spend in its category.
Share of market (SOM)A brand's share of category sales.
Extra share of voice (ESOV)SOV minus SOM. Positive ESOV is linked to market share growth; see the Hive entry on excess share of voice.
Very large business effectThe IPA's measure of a campaign reporting a very large gain in an outcome such as sales, share, price, penetration or profit.
Presence floorThe minimum level of marketing a business commits never to go below.

Cutting, holding and raising compared

Cut to zero (go dark)Hold share of voiceRaise share of voice
Cash this yearBestMiddleWorst
Market share over 1 to 3 yearsFalls steadily[1]Roughly stableTends to grow[2]
Cost of attentionNot relevantOften lower than usualOften lower than usual
Position when demand returnsRebuilding from a weaker baseEvenAhead
When it makes senseSurvival is at stakeMost firms with stable cashFirms with cash and a market opportunity

Chapter 2 · 1 min

The evidence, part 1: what going dark costs

In 2023 the Journal of Advertising Research published a replication and extension by Peilin Phua, Nicole Hartnett, Virginia Beal, Giang Trinh and Rachel Kennedy, mostly of the Ehrenberg-Bass Institute. An earlier study by the same group had looked at alcohol sales in Australia. The 2023 study covered 365 brands in 22 consumer goods categories sold widely in US supermarkets, and tracked brands that stopped advertising for a year or longer.[1]

Figure 1. Brands that go dark keep losing share

10%
Year 1
20%
Year 2
28%
Year 3

Years without advertising; market share lost (%), 0 to 30

Average decline in market share relative to the last year the brand advertised. Source: Phua, Hartnett, Beal, Trinh and Kennedy, "When Brands Go Dark", Journal of Advertising Research, 2023, as summarised by the ARF.

The ARF's summary of the paper says market share "declined on average at a steady rate, year-over-year". It did not level off. And the paper found that "brands already in decline and smaller brands suffered a greater loss of market share."[1]

That second sentence matters more than the averages for most readers. A large brand has years of memories stored in millions of heads. A small brand has a thin layer, and it wears off fast. Byron Sharp's How Brands Grow explains why: much of a brand's sales come from light buyers who barely think about it, and they are the first to forget.

Limits of the going-dark evidence

Even so, the mechanism carries across categories. People forget brands they stop seeing, and new buyers never learn about them at all.

Chapter 3 · 3 min

The evidence, part 2: what the 2008 to 2009 recession showed

During the 2008 to 2009 recession Peter Field wrote a report for the IPA, Advertising in a Downturn. In April 2020 he revisited it for the LinkedIn B2B Institute, examining IPA data from around 50 case studies that advertised through the recession.[2][3] He split them into three equal-sized groups by extra share of voice: zero or less, 0 to 8%, and over 8%. None of the three had gone dark.

Figure 2. Brands that pushed share of voice grew share fastest

1.0%
0% or less
1.4%
0-8%
4.5%
Over 8%

Extra share of voice band; annual market share growth (%), 0 to 5

Annualised market share growth by extra share of voice, IPA cases covering the 2008 to 2009 recession. Source: Field, P., "Advertising in a downturn revisited", IPA, 2020, chart "Investing in SOV drives strong growth during recession".

The group with more than 8% extra share of voice grew market share 4.5 times as fast as the group at zero or below. The same chart shows 2.1 very large business effects on average for that group, against 0.4 for the under-investors. On profit, 38% of the high-investment cases reported very large profit growth, against 10% of the middle group and none of the under-investors.[2]

Field also compared each group's performance in the recession with cases from the two years either side of it.

Figure 3. The same spend works harder in a recession

46
0% or less
88
0-8%
167
Over 8%

Normal times: index 100

Extra share of voice band; index, 0 to 180

Annualised market share growth in the 2008 to 2009 recession against cases from the two years either side, indexed. Source: Field, P., "Advertising in a downturn revisited", IPA, 2020.

Under-investing hurt more than usual, and investing paid more than usual. Field's summary: "Benefits of investment grew strongly during recession compared to normal times", and "Penalties of under-investment worsened".[2]

Limits of the 2008 evidence

An older case: Kellogg in the Great Depression

The story is retold often, so we stick to what one reference work confirms. According to the International Directory of Company Histories, during the Great Depression Kellogg's directors decided to cut advertising, premiums and other expenses. W.K. Kellogg returned from California, called a meeting and told the officers to press ahead, and they voted again, "this time adding $1 million to the advertising budget". The entry says profits improved from around $4.3 million a year in the late 1920s to $5.7 million in the early 1930s.[7] One company proves nothing about cause, but it is the bet in its plainest form.

What Field advises

Field's practical advice follows from the numbers. "Brand advertising is not about profiting in recession, it is about capitalising on recovery," he writes. Where revenues fall far enough to threaten survival, "there may be no choice but to cut all advertising to conserve cash".[3] Outside that case, he recommends defending share of voice.

The same logic holds in B2B. In Binet and Field's B2B cut of the IPA data, 10% extra share of voice was associated with 0.6 percentage points of market share growth a year for consumer brands and 0.7 for B2B brands.[4] The Hive entry on The Long and the Short of It covers the wider research behind these numbers.

If you want a second opinion on which parts of your marketing are carrying the load before you cut anything, the Brave New audit looks at what a buyer sees and which pieces bring in customers.

Chapter 4 · 4 min

Original research: how US advertising moves through recessions

The studies above describe what happens to brands. We wanted to see what happens to the advertising market around them, because that is what makes share of voice cheaper or dearer. Every series below is free and public. We marked recessions with FRED's USREC series, which follows the National Bureau of Economic Research dates, and found four since 1990: August 1990 to March 1991, April to November 2001, January 2008 to June 2009, and March to April 2020.

Finding 1: advertising jobs fall harder than jobs overall

We used the Bureau of Labor Statistics' monthly count of jobs in advertising, public relations and related services (series CES6054180001), the agencies and services that plan, make and place advertising. It is a proxy for advertising activity, not a measure of spending. For each recession we compared the last month before it began with the lowest point in the following three years, and did the same for all US nonfarm jobs.

Figure 4. Advertising jobs fell 2 to 8 times as far as all jobs in three of four recessions

Advertising jobsAll nonfarm jobs
-5.2%
-1.4%
1990-91
-15.1%
-2.0%
2001
-14.3%
-6.3%
2008-09
-10.8%
-14.4%
2020

Change in jobs (%), 0 to -16

Change from the last month before each recession to the lowest month within 36 months of its start, seasonally adjusted. Source: Brave New Media analysis of BLS Current Employment Statistics (CES6054180001, CES0000000001) and FRED USREC, 3 November 2026.

In 1990 to 1991, advertising jobs fell 3.7 times as far as jobs overall. In 2001 the ratio was 7.7 times, and in 2008 to 2009 it was 2.3 times. The 2020 recession was the exception: lockdowns hit service jobs of every kind, and all jobs fell further than advertising jobs. Field had warned that 2020 was "not a 'normal' recession".[2]

The pattern in the normal recessions is the point. When demand falls, advertising contracts faster than the economy. That means the total noise in most categories drops, and every brand still speaking holds a larger share of what is left.

Figure 5. Advertising takes years to recover

RecessionAdvertising and related services jobs (months)All nonfarm jobs (months)
1990-915331
200126246
2008-098177
20202628

Months from the last month before each recession until the series first returned to that level after its low. Source: Brave New Media analysis of BLS Current Employment Statistics, 3 November 2026.

After 1990 to 1991, advertising jobs took 53 months to recover, against 31 for all jobs. After 2001 they took 262 months, until January 2023. That figure mixes the recession with the end of the dot-com boom and a long shift in how advertising is made and bought, so it overstates the recession's own effect. It still shows how far below its old level the industry stayed.

Finding 2: advertising revenue swings harder than GDP

The Census Bureau's Quarterly Services Survey reports revenue for advertising, public relations and related services firms (FRED series REV5418TAXABL144QNSA) from late 2003. We compared each quarter with the same quarter a year earlier, alongside nominal GDP.

Figure 6. Ad services revenue fell four times as far as GDP in 2009

Advertising services revenueNominal GDP
-10010202008 Q12008 Q32009 Q12009 Q32010 Q12020 Q12020 Q32021 Q222.817.5

Year-on-year change (%), -15 to +25. 2008 Q1 to 2010 Q2, then 2020 Q1 to 2021 Q2

QuarterAdvertising services revenue (%)Nominal GDP (%)
2008 Q15.63.5
2008 Q22.13.2
2008 Q36.42.3
2008 Q40.5-0.7
2009 Q1-6.8-1.9
2009 Q2-11.3-3.3
2009 Q3-13.7-3.0
2009 Q4-4.20.3
2010 Q11.02.3
2010 Q210.24.2
2020 Q111.03.0
2020 Q2-9.4-6.7
2020 Q3-0.4-0.1
2020 Q40.30.7
2021 Q1-0.44.4
2021 Q222.817.5

Change on the same quarter a year earlier. Advertising services revenue is for firms subject to federal income tax, not seasonally adjusted. Source: Brave New Media analysis of US Census Bureau Quarterly Services Survey and BEA GDP via FRED, 3 November 2026.

At its worst, in the third quarter of 2009, advertising services revenue was 13.7% below a year earlier. Nominal GDP's worst quarter was a 3.3% fall. In 2020 the gap was smaller, -9.4% against -6.7%, and both bounced back hard in 2021. In both downturns, advertising fell later than the economy and further.

Field's own notes on 2008 to 2009 show where the cuts landed. "Long-term brand media down 15-20%," his chart reads, while "short-term digital media up 20%".[2] Firms did not stop marketing. They moved money from building demand to harvesting it, the start of a drift to short-term activation that Field says "has cost businesses dearly over the years since then".[3]

Finding 3: in 2020, attention got cheaper

The Bureau of Labor Statistics also tracks the prices advertisers pay. We compared the change from February to June 2020 with the average change over the same months in the five years before, to strip out the normal seasonal pattern.

Figure 7. US ad prices fell sharply in spring 2020

2020 (Feb to Jun)2015 to 2019 average (Feb to Jun)
Advertising space and time, all media
-8.1%
0.6%
Search and text ads online
-13.1%
-3.6%
Consumer prices (all items)
-0.9%
0.7%

Price change (%), -14 to +6 (axis extended past +2 so the small positive values stay visible)

Change in producer price indexes (WPU36; PCU519130519130101) and the consumer price index (CPIAUCSL), not seasonally adjusted except CPI. Source: Brave New Media analysis of BLS data via FRED, 3 November 2026.

Advertising space and time fell 8.1% in four months, against a normal rise of 0.6% over the same months. Search and text ads, which usually drift down in spring, fell 13.1% against a normal 3.6%. Consumer prices barely moved. The all-media index was back above its February level by October 2020.

This matches Field's third lesson: "The cost of SOV falls during recession."[2] A business that held its budget flat in spring 2020 bought more advertising than it had in February, in categories where many rivals had gone quiet.

Limits of our analysis

Chapter 5 · 1 min

A framework: the Cut-or-Hold Matrix

The research points to two questions that decide the right move: is survival at stake, and where does your share of voice stand against your share of market? Together they give four positions.

Your positionWhat to doEvidence
Survival at stake, ESOV already at or below zeroCut to a presence floor, never to zero. Keep the cheapest ways buyers find and remember you.Field: cutting all advertising may be unavoidable when survival is at risk;[3] going dark loses share steadily[1]
Survival at stake, ESOV positiveTrim, but stay at or near share of market. Rivals' cuts may let you hold share of voice with less spend.Maintaining SOV "may entail no extra spend";[2] small brands lose most when dark[1]
Cash stable, ESOV at or below zeroHold or raise spend to reach share of market. This is the group that fared worst in 2008 to 2009.ESOV of zero or less: 1.0% share growth a year, index 46 against normal times[2]
Cash stable, ESOV positiveRaise share of voice while it is cheap, weighted to brand building.ESOV over 8%: 4.5% share growth a year, index 167;[2] ad prices fell 8.1% in spring 2020

Most firms do not know their share of voice to the decimal, and in B2B it is hard to measure. A rough answer is enough: are we louder or quieter than our share of the market, and are our rivals cutting?

Chapter 6 · 2 min

How to cut your marketing budget without losing ground

Step 1: Sort every line into three piles

List every marketing cost, from ads and software to trade shows and freelancers. Mark each one "working" (you can connect it to enquiries, sales or repeat orders), "unclear" (it might work, but nobody has checked) or "waste" (nobody can say why you pay for it).

Step 2: Cut the waste first

Cancel unused subscriptions, stop campaigns with no results and drop events you attend from habit. This is the safest money in the budget.

Step 3: Test the "unclear" pile one item at a time

Pause one item and watch enquiries for a few weeks. If nothing changes, cut it. If enquiries drop, it was working. Tip: Change one thing at a time. If you cut five things in a week and sales dip, you will never know which cut caused it.

Step 4: Protect brand building, not only the ads that report fastest

Most cuts go wrong here. Teams keep the ads that bring this month's leads and drop everything that builds their name. Field's 2008 data shows the cost of that shift, and his 2020 article notes that most businesses already spent "less than 50% on brand-building".[3] Keep your look and voice exactly the same while you trim: your distinctive brand assets make every smaller message work harder.

Step 5: Set a presence floor

Decide the minimum you will never go below. For an industrial firm that might be a monthly email to customers and distributors, two useful posts a week, a current product catalogue online, and one trade event a year. Check any cut against mental and physical availability: does it make you harder to think of, or harder to buy from?

A lean team can hold that floor across markets with an AI flow. One approved piece, say a short note on lead times or raw-material prices, becomes the customer email, the posts, a distributor update and a translated version for each market you sell into, with a person checking each. That is how we think about going broad in a downturn: more reach from the same people, not more spend.

What to cut and what to keep

Marketing activityCut it?Why
Unused software and toolsYes, nowPure cost, no effect on buyers
Ads with no tracked resultsPause and testMay be working, may be waste
Heavy discounting and promotionsReduceField found short-term promotional strategies lead to dependency and profit loss[2]
Email to existing customers and distributorsKeepCheap, and they are the easiest sale
Website, catalogue and spec pagesKeep and fixYour shop window when ads stop
Brand building content and PRTrim, never zeroBuilds the memory that makes ads work
Sponsorships with no audience fitYesVisibility to the wrong people

Which to do first

ActionEffortLikely impactWhen
Sort every line into three pilesLowHighThis week
Cut the waste pileLowMediumThis week
Estimate share of voice against share of marketMediumHighThis month
Set a presence floorLowHighThis month
Test the unclear pile one item at a timeMediumMediumOver the quarter
Raise brand spend if cash allowsMedium to highHigh in the recoveryWhen prices fall

Chapter 7 · 1 min

What this means for leaders

For manufacturers and supply-chain businesses, the bill for going dark arrives late. Buying cycles are long. A buyer who stops hearing from you this year is drawing up next year's supplier list without you, and our data shows the advertising market can take years to return to its old level.

The upside is the mirror image. When rivals cut, every unit of your presence counts for more and often costs less. A downturn is one of the few times a mid-sized supplier can gain share of voice against larger rivals without outspending them, and the 2008 data suggests that share sticks into the recovery. That makes the recession budget a growth decision, especially for firms planning to enter new markets once demand returns. The brands that kept talking will already be known there.

If you want an AI flow that keeps you visible in more markets on a smaller budget, start with the Brave New audit.

Chapter 8 · 1 min

Common mistakes

Chapter 9 · 1 min

Methodology

Recession dates. FRED series USREC (NBER-based recession indicator, monthly). A recession runs from the first to the last month marked 1. We studied the four recessions since January 1990.

Jobs. BLS Current Employment Statistics, seasonally adjusted: advertising, public relations and related services (CES6054180001, available from January 1990) and total nonfarm (CES0000000001), downloaded through the BLS public API on 3 November 2026. For each recession we measured the change from the last month before it to the lowest month within 36 months of its start, and the months until the series first returned to its pre-recession level after that low.

Revenue. US Census Bureau Quarterly Services Survey, total revenue for NAICS 5418, establishments subject to federal income tax, not seasonally adjusted (FRED REV5418TAXABL144QNSA, from the fourth quarter of 2003), compared year on year with nominal GDP (FRED GDP).

Prices. BLS producer price indexes for advertising space and time sales (WPU36, from June 2009) and internet search and textual advertising sales (PCU519130519130101, from December 2009), and the consumer price index (CPIAUCSL). We compared February to June 2020 with the average February to June change in 2015 to 2019.

Limits. Jobs and revenue are proxies for advertising activity. The series cover the US only. The price series cover one recession. Data are as published on 3 November 2026 and may be revised.

The data behind every figure, and the script that produced our own, are published with this essay as CSV files.

Chapter 10 · 1 min

Frequently asked questions

Should you cut your marketing budget in a recession?

Cut waste, not presence. Remove spending you cannot connect to customers, test what is unclear, and protect what brings enquiries. Brands that stopped advertising lost on average 10% of market share after one year and 28% after three, with small brands hit hardest. If survival is at stake, cut to a presence floor rather than to zero.

What happens if a business stops advertising?

Market share usually falls, and keeps falling. A 2023 Ehrenberg-Bass study of 365 US supermarket brands found share fell by about 10% after one year without advertising, 20% after two and 28% after three, relative to the last year advertised. Smaller brands and brands already in decline lost the most.

Is a recession a good time to advertise?

Often, yes. In three of the last four US recessions, advertising jobs fell far faster than jobs overall, and in spring 2020 US ad prices fell 8.1% in four months. In Peter Field's analysis of 2008 to 2009 IPA cases, brands with high extra share of voice grew market share 4.5 times as fast as brands at zero or below.

How much should a business spend on marketing during a downturn?

There is no single number. A useful target is to keep share of voice at or above share of market, which may need less money than usual because rivals cut. If cash is tight, set a presence floor you never go below. If cash is stable, consider raising brand spend while attention is cheaper.

What marketing should you cut first in a recession?

Cut unused tools, ads with no tracked results, sponsorships that reach the wrong people and heavy discounting. Keep email to existing customers and distributors, your website and product pages, and some brand building. Cut one thing at a time and watch enquiries, so you know which cuts cost you and which were safe.

Your next step

This week, list every marketing cost you pay and mark each one "working", "unclear" or "waste". Then write one sentence on where your share of voice stands against your share of market, and one on what your main rivals are doing. Those two lines tell you which box of the Cut-or-Hold Matrix you are in. You can see how we approach growth work in our case studies.

If you want a team to keep you present in every market you sell to while you trim, start with the Brave New audit.

Brave New Media is a marketing agency that builds AI flows to help companies scale their ideas, grow their customer base and expand into new markets, with experience across supply chain and manufacturing.

See also

References

1. Phua, P., Hartnett, N., Beal, V., Trinh, G. and Kennedy, R. "When Brands Go Dark: A Replication and Extension. Examining Market Share of Brands that Stop Advertising for a Year or Longer." Journal of Advertising Research, 2023. Summary by the ARF. thearf.org

2. Field, P. "Advertising in a downturn revisited: key learnings from 2008 and new findings in the era of COVID-19." Webinar charts, IPA, April 2020. ipa.co.uk

3. Field, P. "Advertising in Recession: Long, Short, or Dark?" The B2B Institute at LinkedIn, April 2020 (encore post January 2021). business.linkedin.com

4. Binet, L. and Field, P. "The 5 Principles of Growth in B2B Marketing." The B2B Institute at LinkedIn, 2019. news.linkedin.com

5. Federal Reserve Bank of St. Louis. FRED series USREC, REV5418TAXABL144QNSA, GDP, WPU36, PCU519130519130101 and CPIAUCSL. fred.stlouisfed.org

6. US Bureau of Labor Statistics. Current Employment Statistics, series CES6054180001 and CES0000000001. bls.gov

7. "Kellogg Company." International Directory of Company Histories, via Encyclopedia.com. encyclopedia.com

8. Brave New Media. US advertising activity through four recessions, 1990 to 2021, 3 November 2026. Data published with this essay.