Your LinkedIn Ads Measure 30 Days. Your Deals Take 272.

Thirty days against two hundred and seventy two, before you add the energy premium

Start with the documented numbers rather than the folklore. LinkedIn's conversion tracking schema sets postClickAttributionWindowSize to a default of 30 days, with allowed values of only 1, 7 or 30. The view through default is 7 days, same allowed values. LinkedIn's own documentation notes that by setting these fields to something else, the advertiser can narrow the window. Narrow. There is no widening on the standard path.

Dreamdata's 2026 benchmark report, built on more than 66 million sessions and over 3.5 million complete customer journeys between August 2024 and July 2025, puts the average complete B2B customer journey at 272 days, up from 211 the year before. For companies with 250 or more employees it is 326 days. Dreamdata is a LinkedIn marketing partner, which is worth stating, but the sample and method are published and the direction is not in dispute.

Now add the energy premium. Operator and NOC procurement commonly runs 9 to 18 months from first contact to award, before any framework agreement or prequalification cycle that preceded it. Against a 30 day default click window that is a factor of 9 to 18. Under default settings, LinkedIn's reporting is structurally incapable of observing the outcome of most of the deals it influenced.

The consequence is not that the spend does not work. It is that the report in front of you is a report on the small minority of outcomes that happened to resolve quickly, presented with the same confidence as a complete picture. Short cycle outcomes are systematically over represented in it. If you optimise against that report, you optimise your media mix toward whatever produces fast, low value conversions, which in energy usually means students, competitors and consultants downloading a specification.

Projet 54A refinery and petrochemical complex at dusk. This is the asset base energy marketers are selling into, on procurement cycles no advertising platform's attribution window was designed to see.

Very few of LinkedIn's own, and none of them broken out for energy

The most quoted figure in B2B paid social is LinkedIn's claim that Thought Leader Ads deliver a 1.7 times higher click through rate and a 1.6 times higher engagement rate than other single image ad campaigns. Both are asterisked to LinkedIn's own footnote: based on a select group of early pilot testers, 8 February to 25 April 2023, and data may be impacted by early adopter bias and is considered directional. It is a ten week pilot of self selected early adopters, now three and a half years old, which LinkedIn itself labels directional. It has not been withdrawn, but it has been demoted: LinkedIn's current product page carries no numeric performance claim at all.

Document Ads are a cleaner case. LinkedIn publishes no performance statistic for them whatsoever. Every percentage you have seen attached to Document Ads in an agency blog is untraceable to LinkedIn.

The credible numbers are third party. Metadata's 2026 B2B paid media benchmark covers calendar 2025 across 153 B2B advertisers, 57.6 million dollars of measured spend and 211,000 leads, with the LinkedIn cuts resting on 138 advertisers. Its method is published: spend weighted aggregates rather than averages of averages, with a minimum of five advertisers and 50,000 dollars of spend per cut and no single advertiser above half that spend. It released 111 cuts as an anonymised dataset. That level of disclosure is rare enough to be the reason to prefer it.

Its headline LinkedIn figures for 2025 are a cost per lead of 202 dollars, cost per click of 9.39 dollars, cost per mille of 63.19 dollars, click through rate of 0.67 per cent and a click to lead conversion of 5.9 per cent. The distribution is the part to internalise. Across 88 advertisers the tenth percentile cost per lead is 109 dollars, the median is 376 dollars, the seventy fifth is 658 and the ninetieth is 1,340. The 202 dollar headline sits around the twenty sixth percentile. If you plan a budget off it you have planned off the biggest spenders' economics.

By format, Metadata puts Document Ads at 142 dollars per lead, single image at 200, video at 265 and conversation at 362, with Document Ads converting at 11.9 per cent click to lead. By destination, Lead Gen Forms cost 193 dollars per lead against 346 for landing pages, a 1.8 times difference. For a technical audience that will not fill in a long form on a slow corporate site, that is the highest return change available and it is free to test.

And here is the finding that should change how energy marketers read all of this. Neither Metadata nor Dreamdata publishes an energy, utilities, industrial or manufacturing breakout. Metadata segments by channel, format, audience type and company size. Dreamdata segments by company size only. As at October 2026 there is no credible, methodologically disclosed, energy specific LinkedIn advertising benchmark in the public domain. Every number in this section describes a largely software weighted population. Treat them as a direction of travel, not as a target.

A heuristic with a published formula, circulated as a measured constant

The source is Professor John Dawes of the Ehrenberg-Bass Institute, writing in a B2B Report of May 2021 commissioned by LinkedIn's B2B Institute. His claim, verbatim, is that up to 95 per cent of business clients are not in the market for many goods and services at any one time. The caveat in the same paper is almost never quoted: the 95 per cent figure is not meant to be a precise rule.

It is a derivation, not a measurement. Dawes reasons that companies switch providers such as a principal bank or law firm roughly once every five years, so about 20 per cent are in market across a year and about 5 per cent in any quarter. Then he publishes the generalisation, which is the genuinely useful part: with a two year average interpurchase time, about 50 per cent are in market over a year and about 13 per cent in any quarter. The ratio is a direct function of your category's repurchase interval.

That is decisive for energy, because energy categories sit at wildly different points on that curve. An annual chemicals supply agreement, a five yearly turbine major overhaul, a framework renewal and a twenty five year power purchase agreement produce four completely different in market percentages. Running all four off a single 95-5 assumption is not conservative, it is arbitrary. Dawes's own instruction is to calculate it: if you know the average interpurchase time for your category you can readily derive the in market share, and where the data does not exist, survey buyers on how often they buy the category.

The practitioner criticism is real but lands on the circulation rather than the research. David Gyertson, digital director at Transmission, wrote in June 2024 that the figure is a finger in the air guesstimate and is misleadingly labelled and presented as research. He cites no counter data. We could find no peer reviewed academic rebuttal at all. The irony is that Dawes pre empted most of the objection in 2021 by saying it is not a precise rule and publishing the formula to recalculate it.

The companion number, the brand to activation split, deserves the same care. Les Binet and Peter Field's 5 Principles of Growth in B2B Marketing gives 46 per cent brand and 54 per cent activation, with the authors' own caveat that this ratio should not be followed too precisely, rather it is a guiding principle. The B2B sample behind it is fewer than 50 IPA Effectiveness Databank cases, skewed toward the UK, toward larger budgets and toward campaigns successful enough to be entered for an award. The authors write, verbatim, that by necessity their findings are tentative. Use it as a prior to be updated, not a budget formula.

Documents, thought leaders, and the one bidding change that matters

Thought Leader Ads sponsor an individual's organic post. Since March 2024 the person does not have to be an employee: LinkedIn's current page says you can boost thoughts from company employees, industry experts, customers or content creators, with the creator approving or denying each sponsorship request. For an energy supplier that unlocks sponsoring a respected independent engineer, a customer, or an analyst, which is a materially different proposition from boosting your own sales director. The constraints are real: no headline, no introductory text, no call to action button, and posts containing multiple images, polls or a document cannot be promoted at all. So a gated whitepaper post cannot be a Thought Leader Ad.

Document Ads let a specification, a technical note or a field study be read in feed and optionally gated behind a Lead Gen Form, with reporting on what percentage of the document was viewed. On the only credible data they are also the cheapest format per lead. For a sector whose buyers genuinely read documents, this is the obvious default and it is under used.

Sponsored Messaging carries a hard geographic constraint that European energy marketers need to know. Message Ads and Conversation Ads cannot target EU based members. New campaigns were blocked from 15 December 2021 and existing ones stopped delivering on 10 January 2022, following the European Court of Justice ruling of 25 November 2021 that native inbox advertising requires direct marketing consent under the ePrivacy Directive. LinkedIn's own product pages make no mention of this. From June 2026 a Not Interested call to action is also rendered automatically on these formats and advertisers cannot opt out.

The single most important 2026 change for long cycle sellers is a bidding target. From API version 202602, February 2026, lead generation campaigns can optimise toward MAX_QUALIFIED_LEAD, using qualified lead data sent through the Conversions API or CRM data shared from Business Manager. Instead of teaching the algorithm to find people who fill in forms, you teach it to find people your own sales team has marked as qualified. In energy, where form fill volume correlates poorly with pipeline, that is the difference between optimising toward noise and optimising toward revenue.

Two more to note and one to be sceptical about. Creator Marketplace and BrandWorks launched on 10 June 2026, the former in North America alpha only. LinkedIn said on 1 July 2026 that advertisers running five or more ad variants see more than 20 per cent higher click through rates, with no methodology, sample or period disclosed. And Accelerate, LinkedIn's AI campaign automation, has an agency objection on the record worth heeding: clients generally have very specific targeting they want to reach, which is precisely the energy case, where the audience is a named list of operators and job functions rather than a broad category.

Stop asking the platform to prove revenue it cannot see

First, fix what is fixable inside LinkedIn. Move conversion tracking off the Insight Tag default and onto the Conversions API, sending at least a hashed email and the LinkedIn first party tracking identifier together, because only matched events can be used for attribution or optimisation. Set the window to 365 days on the conversion types that support it. Then read the conversions FAQ carefully, because the binding constraint is not the window, it is the rule that the conversion timestamp should be within 90 days. The 365 day attribution window governs how far back LinkedIn looks from a conversion for a click. The 90 day submission cap governs how late you may report the event at all. Both bind, and conflating them is the most common error in this whole area.

Second, check your CRM against the list before promising anything. LinkedIn's Revenue Attribution Report supports HubSpot, Microsoft Dynamics 365 and Salesforce Sales Cloud. That is the complete list. A large share of the energy sector runs on SAP, Oracle or something bespoke, and is therefore excluded from LinkedIn native revenue reporting entirely. Find that out before the budget conversation, not during it.

Third, accept that incrementality testing is priced out of most energy marketing budgets. LinkedIn's brand lift testing is a genuine randomised holdout, splitting the audience into test and control groups and surveying both. It requires a minimum of 45,000 US dollars of budget for a single question, 45,000 more per additional question, and 90,000 dollars for the test duration, and it runs at account level across every campaign. Know the threshold so you can say no to it credibly.

Fourth, and most important, change what you report. On a 272 day journey the quarterly question is not what did it return, it is are we getting closer. Four measures move inside a quarter and are defensible to a finance director: the share of the known buying group reached at each named target account, not just the champion, since Ty Heath puts the average B2B buying group at about 8.2 people; bid list and prequalification presence, which is the only outcome that actually gates revenue in this sector; inbound enquiries from named target accounts regardless of attributed source; and document read depth on technical content, which is the closest thing to a genuine interest signal the platform offers.

Finally, a gap worth naming. In researching this piece we could not find a single directly attributed quote from an in house energy or industrial marketing leader on long cycle measurement anywhere in the public record. The software sector has published its measurement thinking at length. This sector has not. That is an opportunity for whoever goes first, and it is also why every benchmark in this article comes with the caveat that it was measured somewhere else.