Reach is not the point: what Arrow ECS proved about cheap attention in B2B

Arrow ECS bought 454,176 views at £0.0019 each during the seven weeks when a large group of VMware customers were reconsidering their infrastructure. The benchmark for that kind of view is £0.08. The campaign also produced 323 clicks, which is a real number and the wrong one to look at. Attention bought during a moment of genuine buyer uncertainty is worth more than the same attention bought at benchmark price in a quiet month.

What did the campaign actually buy?

Seven weeks of video amplification for Arrow ECS around VMware, running through the Broadcom transition. The full read:

One Shorts asset, "Containers Need Full Stack", took 169,660 views on its own. County Dublin alone accounted for around 121,700. Joe Baguley, VMware's CTO for Field Sales in EMEA, carried the message.

Take those numbers to a media buyer and the reaction is usually the same. The cost per view looks like a reporting error. It is what happens when the market genuinely wants the information you are handing out.

Why is attention cheaper when buyers are unsettled?

Because auction pricing is set by advertiser demand, and audience interest is set by what is happening in the audience's working week. Those two things come apart during a transition.

A large enterprise infrastructure change puts thousands of technical people into an unplanned review at the same time. They did not schedule it. Nobody in that group woke up in January intending to spend their spring reading about container strategy. They are doing it because their supplier position moved and they now have to form a view they can defend to their own board.

At that moment their appetite for a credible explanation is unusually high, and the cost of reaching them has not moved to match, because most competitors are still running the plan they signed off in November. That gap is the whole opportunity, and it closes.

This is the argument for planning against trigger events rather than against the calendar. Acquisitions, licensing changes, regulatory deadlines, end of support dates, a new head of function. Each of them converts a share of the market from not looking to actively looking, on a date you can often see coming months in advance. Professor John Dawes at the Ehrenberg Bass Institute, writing for the LinkedIn B2B Institute in 2021, put the share of B2B buyers in market at any given moment at around 5%. Trigger events are the mechanism by which that share temporarily changes, in a segment you can name.

Most annual plans have no mechanism for this at all. Budget is committed twelve months ahead, in equal monthly slices, and when the moment arrives there is nothing uncommitted to move.

Why did a named engineer outperform brand messaging?

Because the audience was technical, and technical audiences resolve uncertainty by deciding who to believe rather than by comparing claims.

Joe Baguley is VMware's CTO for Field Sales in EMEA. He has a public record, a technical reputation and a face that a large part of that audience already recognised. When he explains what a change means, the viewer is assessing a person they can look up. A brand statement offers nothing to assess. It has no track record, it cannot be cross checked, and it carries an obvious commercial motive.

There is a governance point underneath this that most B2B businesses get wrong. The named expert has to be allowed to say something that costs them. If every sentence has been through three approvals and lands on a marketing message, you have spent your credibility to deliver a claim the audience already discounts. The value of a named human is that they can concede a limitation, and a brand cannot.

Gartner's B2B buying research puts a typical complex purchase in the hands of six to ten people. In a technical purchase, a good proportion of those people are engineers who will never fill in a form, never speak to your sales team, and will privately veto anything they consider unserious. They are the audience for a named expert. They are unreachable by brand advertising.

Why did YouTube Shorts do the heavy lifting?

Because it was the cheapest place to buy a completed view from the right people, and B2B teams have largely left it alone.

"Containers Need Full Stack" ran to 169,660 views as a single asset. The format works for this audience for reasons that have nothing to do with the platform being fashionable. A short vertical video is watched in a gap: between meetings, on a phone, at the point in the day when nobody is going to open a 40 minute webinar recording. It is also the only format where a technical viewer will give you a complete watch, which is a far more honest signal than an impression.

The strategic use of it is as a reach layer underneath something heavier. Shorts buys the volume and the recognition at a price that makes the arithmetic work. The longer form assets, the webinars and the technical content do the persuading for people who have already decided you are worth the time. Running the reach layer alone produces cheap views that go nowhere. Running only the long form produces excellent content nobody finds.

The other reason to use it is that the price will not last. Cheap inventory in B2B is always a temporary condition created by the rest of the market not being there yet.

Is 323 clicks a failure?

By any click based standard, yes. Which is why the click based standard is the wrong instrument here.

Nobody watching a two minute video about container strategy on a phone was going to click through and buy enterprise infrastructure that afternoon. That purchase involves a buying group, a procurement process, an incumbent supplier relationship and a budget cycle. The click has no role in it.

What the campaign was buying was position in a decision that had not started yet. Some months later, a shortlist gets drawn up, and the question in the room is who was talking sensibly when everything was uncertain. That is what 454,176 views at £0.0019 was purchasing, and the analytics record of it will eventually arrive as a direct visit or a branded search from someone your reporting has never seen before.

Judged on clicks, the campaign cost £2.70 per click and looks ordinary. Judged on what it was for, it bought attention roughly 42 times more efficiently than the benchmark, from a named technical audience, in the specific weeks that audience was paying attention. Both readings come from the same dataset. Only one of them describes what happened.

The practical defence against the first reading is to declare the measure before launch. Nominate cost per view and completion as the layer you will judge on, agree it with whoever will ask about clicks, and write it down while the campaign is still a proposal. Deciding what counts after the numbers land is how good campaigns get cancelled.

What did County Dublin tell us?

That reach reports an average, and the average hides the finding.

County Dublin alone accounted for around 121,700 views, out of 454,176. One small geography, a substantial share of the total attention. That concentration reflects where a particular cluster of technical infrastructure people actually sit, and it is the kind of thing you only see if your reporting can break the number down and if the reporting belongs to you.

That last part matters commercially. Arrow's reporting on this sits with Arrow. When measurement lives inside the platform that sold you the media, you get the platform's view of its own contribution, and you lose the history on the day you stop paying. A geographic concentration finding like this one is worth more than the campaign that produced it, because it tells you where to put the next three campaigns, the next event and possibly the next hire.

Where to start this quarter

List the trigger events visible in your market over the next twelve months. Acquisitions, end of support dates, regulatory deadlines, licensing changes. Then look at your budget and see how much of it is uncommitted in those specific weeks. In most plans the answer is nothing, and that is the fixable problem.

Find the one person in your business or your vendor's business who could credibly explain a hard question on camera, and get agreement in principle that they can speak without the approval chain rewriting them.

Then set the measure before you spend. Decide which signal you will judge cheap reach on, at what threshold and after how many days, and put it in writing while everyone is still calm.

Our demand generation programmes are built to run against trigger events rather than the calendar, and our conversion and attribution intelligence work is what stops a campaign like this being judged on the wrong column.

Frequently asked questions

It depends entirely on audience and timing. The working benchmark used on the Arrow ECS campaign was £0.08 per view, and the campaign delivered £0.0019 by running during the Broadcom transition when the target audience had an unusual appetite for the subject. Treat published benchmarks as a starting reference and judge your own cost per view against your own history.

As a cheap reach layer, yes. On the Arrow ECS campaign a single Shorts asset took 169,660 views. Short vertical video gets watched in gaps in the working day and produces completed views rather than impressions. Use it to buy volume and recognition underneath longer form content, rather than expecting it to carry a technical argument on its own.

Choose the signal before launch. Cost per view, completion rate and view volume against a floor are readable within days. Qualification signals arrive at 14 to 30 days and pipeline at 60 to 180. Agree which layer the campaign will be judged on while it is still a proposal, because deciding after the data lands invites the wrong comparison.

Technical audiences resolve uncertainty by deciding who to believe. A named person with a public record can be checked, and can concede a limitation, which a brand statement cannot. On the Arrow ECS campaign the message was carried by Joe Baguley, VMware's CTO for Field Sales in EMEA, during a period when the audience's supplier position was in question.

Something that moves a share of your market from not looking to actively looking on a date you can anticipate: an acquisition, a licensing change, an end of support date, a regulatory deadline, a new head of function. Attention is unusually cheap in those weeks because audience interest rises before competitor spending does.