The best empirical paper on the sales and marketing interface concludes that some differences between the two functions improve performance. Not tolerate. Improve. That finding is the opposite of what every alignment deck asserts, it comes from a survey of 337 business units published in a top journal, and it has been in print since 2007.

Meanwhile the statistics everyone quotes fall apart on contact with their sources: one is definitionally circular, one does not exist in the document it cites, one is a unit error, and one traces to a trade magazine in 2001. Worth going through both halves.

What the research found: two kinds of difference

The study surveyed strategic business units across seven sectors, screening by telephone to confirm each one had both a marketing subunit and a sales subunit. It contacted 1,700 executives and obtained 337 usable questionnaires, a 20% response rate. Self-reported market performance was checked against independent three-year return-on-sales data for 185 of the 337 firms.

It measured five ways the two functions can differ, and estimated a separate model for each.

Every one of the five differences damaged cooperation. Time-horizon differences at -.23, interpersonal skills at -.28, customer-versus-product orientation at -.11, product knowledge at -.08, market knowledge at -.07. And cooperation quality reliably helped performance, at +.21 to +.25 across the five models.

But the direct effects split cleanly by type of difference:

Difference between marketing and salesOn cooperationDirect, on market performance
Customer vs. product orientation-.11+.14
Short-term vs. long-term orientation-.23+.10
Market knowledge-.07.00
Product knowledge-.08-.09
Interpersonal skills-.28-.13

The authors state it plainly: “the signs for different competences between M & S are opposite to the signs for different orientations.”

Netting the cooperation penalty against the direct effect, orientation differences come out positive overall: +.11 for customer-versus-product orientation, +.04 for time horizon. Competence differences come out negative.

The practical reading is precise and it is not what alignment programs usually do. Make the two functions equally competent. Let them see the market differently. Sales leaning toward the customer and the near term, marketing toward the product and the long term, is not a defect to be trained out; the data say it pays, and it pays more than the cooperation friction it causes. What costs you is one function being worse than the other at product knowledge or at handling people.

The authors declare their limits: a fairly low response rate, meaning respondents may already be unusually concerned with this interface; a single informant answering for both functions, with the common-method bias that implies; nearly half of respondents having worked only one side, so their view of the other is an outsider’s; and marketing and sales units that “are not homogeneous” across organizations because companies use the words differently.

Effects of five differences between the marketing and sales functions on cooperation quality and on market performanceEffects of five differences between the marketing and sales functions on cooperation quality and on market performance, from a survey of three hundred and thirty seven strategic business units estimated as five separate structural models. All five differences reduce the quality of cooperation between the two functions. The difference in customer versus product orientation has a path coefficient of minus zero point one one on cooperation, the difference in short term versus long term orientation minus zero point two three, the difference in market knowledge minus zero point zero seven, the difference in product knowledge minus zero point zero eight, and the difference in interpersonal skills minus zero point two eight. Cooperation quality in turn raises market performance in all five models, with coefficients of plus zero point two four, plus zero point two five, plus zero point two five, plus zero point two four and plus zero point two one, all significant at the one percent level. The direct effects of the differences on market performance, however, split by type. Differences in orientation raise performance directly: plus zero point one four for customer versus product orientation, and plus zero point one zero for time horizon. Differences in competence lower it: zero point zero zero and not significant for market knowledge, minus zero point zero nine for product knowledge, and minus zero point one three for interpersonal skills. Netting the indirect cooperation penalty against the direct effect gives total effects of plus zero point one one for customer versus product orientation and plus zero point zero four for time horizon, so orientation differences are net positive despite damaging cooperation. The authors summarise this by stating that the signs for different competences between marketing and sales are opposite to the signs for different orientations. Descriptively, sales inclined toward customer orientation and toward the short term, marketing toward product orientation and the long term, and sales showed higher market knowledge, all significant at the one tenth of one percent level, while the two functions did not differ significantly on product knowledge or on interpersonal skills. Model fit across the five models showed chi square over degrees of freedom between two point seven and three point three, root mean square error of approximation between zero point zero eight and zero point zero nine, and goodness of fit index between zero point nine five and zero point nine six.Two kinds of difference, opposite signsDifference between the two functionsOn cooperationOn performanceCustomer vs. product orientation-.11+.14Short-term vs. long-term horizon-.23+.10Market knowledge-.07.00Product knowledge-.08-.09Interpersonal skills-.28-.13Make them equally competent. Let them see the market differently. That is what the coefficients say.
All five differences hurt cooperation. Only the competence differences hurt performance. Source : Homburg and Jensen, The Thought Worlds of Marketing and Sales: Which Differences Make a Difference?, Journal of Marketing 71(3), 2007, Table 2 (2007)

What does not exist

There is no meta-analysis of sales and marketing integration and performance. None. The nearest thing is a meta-analysis of a different construct, market orientation, pooling 418 effects from 130 independent samples in 114 studies. It reports a market-orientation-to-performance correlation of .32, and, more usefully here, an interdepartmental-connectedness-to-market-orientation correlation of .56 and an interdepartmental-conflict correlation of -.28.

That paper also contains the single most useful methodological fact in this whole area. Among the moderators it tested, the relationship is stronger in studies using subjective, self-reported performance measures than in studies using objective ones. Every alignment statistic in circulation is self-reported. That moderator predicts the direction of the inflation before you even open the sources.

The other adjacent meta-analysis, on cross-functional integration and new product success across 25 studies and 146 correlations, is deliberately hedged: “though cross-functional integration may indeed have a direct impact on success, the combination of integration with other variables may be of greater importance”, and other variables affecting the relationship “reflect researchers’ methodological decisions, suggesting that care should be taken when designing and interpreting the results of such studies.”

There is also no agreed definition. A 2005 conceptual paper exists precisely because integration, alignment, involvement, collaboration and communication get used interchangeably. That paper, incidentally, is often cited as evidence that alignment drives growth. It is a set of propositions with no sample and no test, and it says so.

Findings of the nearest available meta analyses and the moderator affecting every self reported alignment statisticFindings of the nearest available meta analyses to the question of sales and marketing alignment, and the moderator affecting every self reported alignment statistic. No meta analysis of sales and marketing integration and firm performance exists. The nearest is a meta analysis of market orientation, a related but distinct construct, pooling four hundred and eighteen effects from one hundred and thirty independent samples reported in one hundred and fourteen studies, with a search closed in June two thousand four, an inclusion rate of sixty one percent and interjudge reliability between zero point nine one and one point zero on a thirty five percent recode. It reports a correlation between market orientation and performance of zero point three two, rising to zero point four six for overall business performance, with zero point two seven for profits, zero point two six for sales and zero point three one for market share. The dimension closest to alignment is interdepartmental connectedness, whose correlation with market orientation is zero point five six across twenty samples and three thousand two hundred and eighty two observations, against a correlation of minus zero point two eight for interdepartmental conflict across four samples and five hundred and thirty observations; in the structural model the connectedness path is zero point three six with a t statistic of three point two seven and the conflict path is minus zero point one four with a t statistic of minus one point five zero. Its moderator analysis reports that the market orientation to performance relationship is stronger in manufacturing than in services, stronger in low power distance and low uncertainty avoidance cultures, and, most importantly for reading practitioner statistics, stronger in studies using subjective self reported performance measures than in studies using objective ones. A second meta analysis, covering cross functional integration and new product success across twenty five studies and one hundred and forty six correlations with twelve moderators tested, concludes that although cross functional integration may have a direct impact on success, the combination of integration with other variables may be of greater importance, and that other variables affecting the integration success relationship reflect researchers’ methodological decisions, so care should be taken when designing and interpreting such studies.The nearest thing to a meta-analysisOn sales and marketing integration: no meta-analysis exists.On market orientation, a different construct: 418 effects, 130 samples, 114 studiesMarket orientation, to performancer = .32Interdepartmental connectedness, to market orientationr = .56Interdepartmental conflict, to market orientationr = -.28The moderator to rememberEffects run larger with subjective, self-reported outcomes than with objective ones.
No meta-analysis of alignment exists. The nearest one measures a different construct, and warns about self-reported outcomes. Source : Kirca, Jayachandran and Bearden, Market Orientation: A Meta-Analytic Review, Journal of Marketing 69(2), 2005; Troy, Hirunyawipada and Paswan, Journal of Marketing 72(6), 2008 (2005)

The four statistics, and where each one breaks

“Aligned companies grow 20% a year, unaligned ones decline 4%.” The figures come from analyst benchmark reports, and the problem is not the sample. It is the sorting. Respondents are classified into Best-in-Class (top 20%), Industry Average (middle 50%) and Laggards (bottom 30%) by aggregate performance score, in which revenue growth is one of the scoring metrics. A surviving report from the same series defines its top tier as, among other things, “14.9% average year-over-year increase in corporate revenue”.

So the fast growers grow faster than the slow growers, which is how they were selected, and they also report more alignment. Presenting that as “alignment produces growth” reverses the arrow. Add a self-selected online panel, self-reported financials, no sampling frame, no response rate, no controls, and primary documents that are no longer publicly retrievable.

“Aligned organizations generate 208% more revenue from marketing.” This one is worse: the number is not in the source. The document it traces to is a 2005 benchmark report with 1,390 responses from 84 countries. Its actual finding is that aligned businesses generate 24% of total revenue from marketing leads against 15% for non-aligned. That is 1.6x, not 3.08x. The figure 208% does not appear in the report and cannot be derived from any pair of numbers in it.

“36% higher customer retention.” A unit error, from the same report. Churn was 7% for aligned businesses and 11% for poorly aligned ones. The reduction in churn is 4/11, or 36%. Retention went from 89% to 93%: an improvement of 4.5%, restated as 36% by treating a churn reduction as a retention gain.

“38% higher win rates.” This one survives the arithmetic. The report gives close rates of 54% for aligned and 39% for non-aligned businesses, and 54/39 is 1.385. What it does not survive is scrutiny of method: the authors write that “the data collected is the opinion of the participants” and ask readers to “use it to inform and influence your own strategy, rather than to accept the findings as ‘prescribed’ changes”. No controls, no significance testing, no sampling frame, and the report is twenty-one years old and routinely re-dated in modern citations.

A fifth figure deserves a note. The widely quoted “24% faster growth and 27% faster profit over three years” appears in a 2012 blog post attributed to a research firm with no report name and no date. The firm’s own co-founder, describing the underlying work three years later, gave different numbers: “up to 19 percent faster revenue growth, and 15 percent higher profitability”. In the same talk he attributed 48% to 79% of B2B growth to market growth, conditions outside the firm entirely, and only 5% to 36% to cross-functional alignment. A range that wide is barely a finding.

Decomposition of business to business revenue growth presented by a research firm and the ranges it attributed to each sourceDecomposition of business to business revenue growth presented by a research firm, and the ranges it attributed to each source. In a two thousand fifteen conference keynote, a co founder of the firm described an analysis of benchmark data for four hundred business to business organizations covering two thousand six to two thousand fourteen, together with responses from three hundred sales, marketing and product leaders. The published summary of that talk stated that organizations maintaining a particular focus achieve up to nineteen percent faster revenue growth and fifteen percent higher profitability than other companies, using the qualifier up to. Those figures differ from the twenty four percent faster growth and twenty seven percent higher profit over three years that circulate under the same firm’s name, which appear in a two thousand twelve blog post attributed only to a study with no report name and no date, three years before the analysis described in the keynote. Neither figure pair has a retrievable report behind it, and the firm was acquired in two thousand nineteen with its research placed behind a subscription and its public pages removed. More usefully, the same keynote presented a decomposition of where business to business growth comes from, attributing forty eight to seventy nine percent to market growth, meaning conditions outside the firm entirely, twelve to thirty four percent to competitiveness and efficiency, and only five to thirty six percent to cross functional alignment. The alignment range spans a factor of seven, which is too wide to support planning, and the largest single component of growth in the firm’s own analysis lies outside the company’s control.Where the same firm said growth comes fromMarket growth, outside the firm48-79%Competitiveness and efficiency12-34%Cross-functional alignment5-36%A range spanning a factor of seven is not a planning input.And the firm’s own published summary said “up to 19 percent”, not the 24% that circulates.
The firm's own decomposition puts most growth outside the company entirely, and alignment in a range too wide to plan against. Source : SiriusDecisions Summit 2015 keynote, John Neeson, reported on siriusdecisions.com (archived); analysis of benchmark data for 400 B2B organizations, 2006 to 2014 (2015)
Four widely quoted sales and marketing alignment statistics and the result of checking each against its sourceFour widely quoted sales and marketing alignment statistics and the result of checking each against its source. The claim that aligned companies grow twenty percent annually while unaligned companies decline four percent comes from analyst benchmark reports in which respondents are classified into best in class, industry average and laggard tiers by aggregate performance score, and revenue growth is itself one of the metrics used to compute that score; a surviving report in the same series defines its top tier partly as achieving a fourteen point nine percent average year over year increase in corporate revenue. The comparison is therefore definitionally circular: the fast growing group grows faster because that is how it was selected. The underlying reports also rely on a self selected online panel with self reported financials, no sampling frame, no response rate and no controls, and are no longer publicly retrievable. The claim that aligned organizations generate two hundred and eight percent more revenue from marketing does not appear in the two thousand five benchmark report to which it is attributed and cannot be derived from any pair of figures in it; that report, based on one thousand three hundred and ninety responses from businesses in eighty four countries, states that aligned businesses generate twenty four percent of total revenue from marketing leads against fifteen percent for non aligned businesses, a ratio of one point six rather than three point zero eight. The claim of thirty six percent higher customer retention is a unit error drawn from the same report: churn was seven percent for aligned businesses and eleven percent for poorly aligned ones, so the reduction in churn is thirty six percent, but retention rose only from eighty nine percent to ninety three percent, an improvement of four and a half percent. The claim of thirty eight percent higher win rates checks out arithmetically, since the report gives close rates of fifty four percent for aligned businesses and thirty nine percent for non aligned, but the report itself states that the data collected is the opinion of the participants and asks readers to use it to inform strategy rather than to accept the findings as prescribed changes, with no controls, no significance testing and no sampling frame, and it dates from April two thousand five. A fifth claim, of twenty four percent faster growth and twenty seven percent faster profit over three years, appears in a two thousand twelve blog post attributed to a research firm without a report name or date, whereas that firm’s co founder, describing the underlying analysis of benchmark data from four hundred business to business organizations three years later, stated up to nineteen percent faster revenue growth and fifteen percent higher profitability, and in the same talk attributed forty eight to seventy nine percent of business to business growth to market growth and only five to thirty six percent to cross functional alignment.Four statistics, checked against their sources”20% growth vs. 4% decline”Circular. The tiers are sorted by revenue growth.”208% more revenue from marketing”Not in the source. The reported ratio is 1.6x.”36% higher retention”Unit error. Churn fell 11% to 7%; retention rose 4.5%.“38% higher win rates”Arithmetic holds: 54% vs 39% close rates.But: “the data collected is the opinion of the participants.”And meta-analysis finds self-reported performance measures produce systematically larger effects than objective ones.
One is circular, one is absent from its own source, one is a unit error, one survives arithmetic but not method. Source : Aberdeen Group benchmark methodology; MarketingProfs and MathMarketing, Benchmark Report: Sales and Marketing Alignment, April 2005; SiriusDecisions Summit 2015 keynote (2005)

The 70% that nobody measured

“Sales never follows up on 70% of marketing leads” is the most cited number in this field, and its trail is instructive because it ends in a trade magazine.

The earliest primary appearance is an August 2001 article: “a Cahners CARR Research report revealed that most marketers’ secret nightmares are true, on average 70% of the leads marketers sent to sales were never contacted at all.” No sample size, no date, no methodology. Cahners was a B2B trade publisher, which is to say a party with a commercial interest in showing that advertising-generated leads were being wasted.

The peer-reviewed paper that made the figure famous did not measure it either. It cites two consultancy reports, one from 2002 and one from 2009, neither published with a documented method. The phrase “sales lead black hole” itself is credited to a blog post. And the paper’s own footnote adds the caveat that everyone drops: the volume of leads “can be so large that it is not possible for the sales reps to pursue all of them, even if they desired to do so; thus, the 70% figure reflects inadequacies of marketing as well as the sales function.”

So a Journal of Marketing paper became the citation of record for a number that paper inherited from trade press. That is as clean an example of laundering as this field offers.

What that paper did measure is better than the number it borrowed. It surveyed sales reps at four B2B firms, 562 returned surveys at a 21.1% response rate, 461 retained for analysis, modeling the proportion of monthly time reps allocate to marketing leads:

  • Perceived prequalification quality: elasticity +1.12. A 1% improvement in how well reps think leads are qualified buys 1.12% more follow-up time. Across the four firms it ranged from .58 to 1.64.
  • Managerial tracking of follow-up: average elasticity -.14. Monitoring whether reps follow up makes them follow up less, on average. It varied by firm from negative to +.33, positive only at the firm with the least experienced reps.
  • Lead volume: elasticity -.07, which the authors describe as “of little managerial concern”.

And the descriptive statistic that carries the story: mean perceived prequalification quality was 2.87 on a 1-to-7 scale. Not a follow-up problem. A lead-quality problem, as perceived by the people asked to work the leads. Lead quality is decided upstream, by the offer, the page and the targeting together, which is why it is hard to move when those three sit with different suppliers instead of under one roof.

Their closing line is the honest summary of the whole literature: “For too long, sales and marketing have blamed each other for the sales lead black hole. There is more than enough blame to go around.”

Determinants of sales representatives’ follow up of marketing leads and the provenance of the seventy percent claimDeterminants of sales representatives’ follow up of marketing leads, and the provenance of the widely quoted claim that seventy percent of marketing leads are never contacted. On provenance, the earliest primary appearance of the seventy percent figure is an August two thousand one trade press article citing an advertising research report by a business to business trade publisher, with no sample size, no date and no methodology given. The peer reviewed paper that popularised the figure cited two consultancy reports, from two thousand two and two thousand nine, neither published with a documented method, and did not itself measure the figure; the paper’s own footnote observes that the number of marketing leads provided to representatives can be so large that it is not possible to pursue all of them even if the representatives desired to do so, so the seventy percent figure reflects inadequacies of marketing as well as of the sales function. What that paper did measure came from surveys of sales representatives at four business to business firms in scientific instruments, chemicals, copiers and computers, with sales forces from fifty five to two thousand five hundred representatives, five hundred and sixty two surveys returned at a twenty one point zero eight percent response rate, five hundred usable and four hundred and sixty one retained after removing representatives reporting zero time on marketing leads. Using Bayesian estimation, it modelled the proportion of monthly time representatives allocate to marketing leads. Perceived prequalification quality had an elasticity of plus one point one two, meaning a one percent improvement in how well representatives perceive leads to be qualified buys one point one two percent more follow up time, with a range across the four firms from zero point five eight to one point six four. Managerial tracking of follow up had an average elasticity of minus zero point one four, making it counterproductive on average, though it varied by firm from negative at two firms to plus zero point three three at the firm with the least experienced representatives, averaging eight point eight three years against eleven point nine two years overall. Lead volume had an elasticity of only minus zero point zero seven, which the authors describe as of little managerial concern. Mean perceived prequalification quality was two point eight seven on a one to seven scale with a standard deviation of one point two five. The authors declare limitations including that all four firms were large and long established, so conclusions might not apply to smaller or younger firms, that both key explanatory variables were perceptual and therefore subject to measurement error and possible psychological bias, and that the dependent variable is a proportion of time rather than hours or lead counts.What moves follow-up, measuredPerceived lead prequalification quality+1.12Lead volume-0.07Managerial tracking of follow-up-0.14The descriptive statistic that carries the storyMean perceived prequalification quality: 2.87 on a 1-to-7 scale.And the 70% itself: a 2001 trade article citing an undocumented report by a trade publisher.”There is more than enough blame to go around.”
Perceived lead quality has an elasticity above 1. Monitoring follow-up has a negative one. Source : Sabnis, Chatterjee, Grewal and Lilien, The Sales Lead Black Hole: On Sales Reps' Follow-Up of Marketing Leads, Journal of Marketing 77(1), 2013 (2013)

What to do with this

Stop buying the growth argument. No causal study establishes that alignment produces revenue. If someone quotes a percentage, ask which document it is in, and check whether the tiers being compared were sorted on the outcome.

Fix competence gaps, not perspective gaps. The measured harm comes from one function knowing the product less well or handling people less well than the other. The perspective difference, sales toward the customer and the near term, marketing toward the product and the long term, tested positive on net. Training that away removes something the data say is worth having.

Work on perceived lead quality before you work on follow-up compliance. Elasticity above 1 on one, negative on the other. The mechanism is not mysterious: reps ration attention by what they expect to be worth working, and telling them you are watching does not change that expectation.

Treat every self-reported alignment statistic as inflated in a known direction. Meta-analysis found subjective performance measures produce larger effects than objective ones. Every number in this area is subjective.

And measure your own interface. Ask your reps what proportion of marketing leads they think are properly qualified, on a scale. If the answer is near 2.87, you have found the constraint, and it is not a process problem.