From £760 to £72: what changed in the HMS Networks lead engine

Cost per qualified lead at HMS Networks fell from around £760 through trade events to around £72. Four changes did it: separating the work that creates demand from the work that captures it, localising properly across seven markets, moving measurement onto infrastructure HMS owns, and reaching engineers in operational technology rather than a generic decision maker. None of the four was a bidding decision.

What did the starting position look like?

HMS Networks sells industrial communication hardware through the Anybus, Ewon, Red Lion and N-Tron brands. Gateways, remote access, industrial switches. The buyer is an engineer in operational technology, specifying a protocol converter for a production line that cannot be taken offline to find out whether it works.

Trade events were the historic engine, as they are across most of industrial manufacturing. Cost per qualified lead through that route sat at around £760.

Worth saying plainly: £760 is a defensible number. It covers a stand, a build, freight, a team out of the business for a week and the cost of following up, divided by the enquiries that survived qualification. Most industrial businesses that run the same calculation honestly land somewhere near it. The larger problem is that most never run it at all, because the exhibition budget sits in a different line, gets approved a year ahead, and never gets divided by anything.

The four changes below are what made the comparison meaningful. Buying cheaper clicks was never going to do it on its own.

First change: separating demand creation from demand capture

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%. In industrial hardware that number feels generous. A gateway gets specified when a line is being built, upgraded or fixed, and those events are not evenly distributed across a year.

That gives you two audiences with nothing in common except the eventual purchase. The engineer already searching for a device that speaks two protocols is in market this month and will find you or a competitor before the end of it. The engineer who has a problem and does not yet know the category exists is everyone else.

Capture work competes for the first group against every rival, at the same time, which is why cost per lead on those channels climbs every year regardless of how well the account is run. Creation work adds to the pool that later arrives on the capture channels already knowing who you are, and arrives cheaper as a result.

In practice this meant two streams of work with different content, different measurement and different clocks. Capture activity read on intent signals within a week. Creation activity was judged on qualification and pipeline across months, which is the only fair way to judge it.

Second change: localisation across seven markets, done properly

The programme ran across seven markets, and translation would not have been enough.

Industrial engineers do not search in general language. They search protocol names. In Europe, PROFINET. In the US, EtherNet/IP. Same underlying problem, different word, different installed base, different set of adjacent standards and different set of things that will already be on the wall of the plant.

A page translated cleanly into German that leads on EtherNet/IP is invisible to the German engineer looking for it, and the reason has nothing to do with the quality of the translation. The term is simply wrong for the market. The reverse is equally true in Ohio. Localisation in industrial marketing means rebuilding the page around the standard the local buyer actually runs, then translating.

There is a second effect that turns up in the cost. Specific technical terms usually carry less competition than generic category terms, because fewer businesses bid on them and the people typing them are further along. Localising to the protocol moves you out of a broad, expensive auction and into a narrow one full of people who know exactly what they need.

Of the four changes, this was the least expensive to execute and one of the most productive. It is also the one most often skipped, because it looks like a translation job on a project plan and gets given to whoever handles translation.

Third change: measurement on infrastructure HMS owns

Measurement runs on infrastructure HMS owns rather than a rented platform, using Matomo alongside CRM data.

Three things that buys you.

One number to argue from. Run three advertising platforms and they will routinely claim more conversions between them than you actually had, because each one is scoring its own contribution with a ruler it made itself. An owned layer above them ends that argument by outranking it.

The history stays. Stop paying a platform and the record leaves with it. In a business with a long cycle, the log is the asset, because it is the only thing that lets you check whether the early signal you acted on in March predicted the revenue that landed in August.

You can count the signals nobody else counts. The technical documentation download. The compatibility page visit. The configuration guide read three times by the same organisation. In industrial buying, these are stronger indicators than a contact form, and they are invisible to most rented setups.

The discipline that came with it mattered as much as the tooling. Two lead numbers were reported side by side and never blended: 1,588 sales ready leads from the website, and 201 confirmed leads into CRM, made up of 157 from the HMS own brands and 44 from other divisions. They measure different things at different points, and keeping both is what makes the £72 figure defensible rather than flattering.

You cannot reduce a cost per qualified lead you cannot see honestly. Owning the measurement is what turned this from a claim into a number a finance director could take apart.

Fourth change: reaching engineers, not decision makers

Gartner's B2B buying research puts a typical complex purchase in the hands of six to ten people. Most B2B targeting is built around one of them, usually the most senior, on the assumption that seniority and influence are the same thing.

In industrial hardware they are not. The engineer who specifies your device may have no budget authority, no purchasing title and no decision maker flag against them in any dataset you can buy. The purchase does not happen without them, and by the time it reaches somebody with signing authority the technical choice has usually been made.

So the targeting runs on problem rather than title. The person looking for a protocol conversion answer at four in the afternoon because a line is down is the person worth reaching, whatever their job title says.

That has consequences for the content. It means technical material at a level of detail that would bore a general business audience: compatibility tables, wiring, configuration, standards, failure modes. Most content marketing advice tells you to write for the buyer. In industrial B2B, write for the specifier, and let the buyer read the case study.

What do trade events still do well?

Five things, and none of them are replaceable by a landing page.

Existing customers walk up and tell you what is broken. There is no digital equivalent of an engineer who has run your kit for six years explaining, unprompted, what he would change. That conversation is worth the stand on its own.

Competitor intelligence compresses. A week on a show floor tells you what four competitors are launching, pricing and worrying about. Assembling the same picture from a desk takes a quarter and is less reliable.

Hardware gets held. For products sold on reliability, putting the thing in an engineer's hand does work no page can do. Weight, build quality and terminal design communicate faster than any specification sheet.

Partner and distributor relationships get maintained. In channel businesses a great deal of the year's relationship work happens in three days, in person, and the cost of not doing it shows up somewhere no lead report will ever pick it up.

A meeting with a buyer already in a cycle beats almost any digital touch. Events produce those meetings, reliably, in a concentrated period.

The mistake is almost always in the arithmetic. Badge scans get counted as leads. The cost never gets divided by the enquiries that survived qualification. Internal time and travel stay out of the calculation entirely. And a trade event lead, gathered in a twenty minute conversation with a product in someone's hands, is a different object from a gated download, so comparing the two without saying so is how an unexamined number gets defended for a decade.

The sensible position is to keep the events, fund them from a smaller share of the budget, brief the stand team to log qualified conversations into the same system as everything else, and calculate the real cost per qualified lead once a year with the internal time included. Do that and events stop being the whole engine and become a good part of one.

What did the four changes produce?

A 35:1 return on media. Cost per qualified lead down from around £760 to around £72. 201 confirmed leads into CRM and 1,588 sales ready leads from the website, across seven localised markets.

The £72 is the number people quote. The one that matters more is that HMS can now show which activity produced which lead, on infrastructure that will still hold the record in three years.

Where to start

Three things, and the first will be uncomfortable.

Calculate your real cost per qualified lead by channel, with internal time and travel included, and with "qualified" defined in writing with sales before you start counting. Most industrial businesses have never put the exhibition line and the digital line in the same table using the same definition.

Pull the actual search terms your buyers use in each market, market by market. If your international pages are translations of English pages, there is a reasonable chance you are invisible on the terms local engineers type.

Move one report onto infrastructure you own. One is enough to begin with. The value is not in the tool, it is in the fact that the history stops belonging to somebody who can switch it off.

Our B2B lead generation work starts with that cost per qualified lead calculation, and our demand generation programmes are built on the split between creating and capturing described above.

Frequently asked questions

What is a good cost per qualified lead in B2B? There is no figure worth quoting across industries, because a qualified lead in industrial hardware and a qualified lead in software are different objects bought at different prices. The number that matters is your own, by channel, calculated with internal time included and with a definition of qualified that sales has agreed in writing. HMS Networks moved from around £760 to around £72 on that basis.

How do you calculate cost per lead for a trade show? Take the full cost, including stand, build, freight, travel, accommodation and the salary cost of everyone out of the business for the week, then divide it by the enquiries that survived qualification rather than by badge scans. Most businesses divide by scans, which produces a number three or four times better than the real one and makes any comparison with digital channels meaningless.

Are trade shows worth it for industrial businesses? Usually yes, for things a website cannot do: customers telling you what is broken, competitor intelligence, putting hardware in an engineer's hands, and maintaining distributor relationships. The problem is rarely the event itself. It is funding events as the whole lead engine, counting scans as leads, and never dividing the real cost by the enquiries that qualified.

How do you reduce cost per qualified lead? Separate the spend that creates demand from the spend that captures it, because piling more budget into capture channels raises the price of every additional lead. Then fix targeting so you reach the person who specifies rather than the person with the most senior title, localise properly for each market, and measure on infrastructure you own so you can see which activity produced qualified people.

What does localisation mean in B2B marketing? More than translation. It means rebuilding content around the standards, protocols and terms the local buyer actually uses, then translating. HMS Networks ran across seven markets where European engineers search PROFINET and American engineers search EtherNet/IP for the same underlying problem. A perfectly translated page built on the wrong standard is invisible to the person looking for it.