​

Read time: 4 mins 58 secs

Presented by ElevenLabs

Production shouldn't decide how many ads you test

The testing plan says every hook needs three tones, for four markets. In reality, most of that plan gets cut before launch due to capacity.

But with ElevenLabs' MCP connected, you type your request into Claude and multiple batches of content come back in one run. Once a winner shows up, Ads Engine can connect to your Google, Meta and LinkedIn ad accounts and localises it into 50+ languages.

Brand Chemistry readers get 10,000 free credits.

Thank you for supporting our sponsors, who keep this newsletter free.

Interested in sponsoring these emails? See our partnership options here.

Clay’s “secret weapon”

In 2024, RevenueHero filled in the demo request form on 1,000 B2B SaaS websites and waited to see who would get back to them. 365 did, with the average reply taking 1 day, 5 hours and 17 minutes, and 60% of them were automated. They published their findings as a report, and you'll see those numbers quoted on benchmark pages from at least 8 other companies.

This year Clay ran the same experiment on more than 6,000 companies. The results all went into a database, and on top of it they spun up an interactive tool on a platform called Ploy, that gives you your own company's results. The research method was the same both times, find the demo form, fill it in, and analyse what comes back. What changed is what they did with the answers, and it’s part of a bigger transition in the perceived value of information.

Two major changes

A downloadable PDF used to be enough for people to give you their email address because the information inside it was inherently valuable, since it was hard to get anywhere else. That isn't true anymore. AI can give all of us things like industry averages in seconds, including from RevenueHero's own report findings.

So then it becomes a question of why people wanted that information in the first place, understanding what they were actually trying to achieve. You read a benchmark to self-assess where you sit, to find out if you have a problem or if you're ahead of the curve. A static report makes you walk that route yourself, find your own figures and compare the two to decide whether it's something you need to act on. Clay's tool skipped to the end and handed each company its own results and suggested actions for improving them, which is exactly the thing they were actually interested in.

Interactive tools have been working as lead magnets for nearly twenty years. HubSpot launched their Website Grader in 2007, and the shape of it has become pretty standard. Enter your URL or email address and get back a personalised score.

But those tools were hard to build. A report only needs a writer and a designer. Something like the website grader needs a database of information, a website page that personalises those numbers for each visitor, and a design system to make sure it's all a seamless and recognisable experience. You inherently needed technical support to do it. I don't know about you, but in my experience, marketing teams asking for engineering support very rarely get it.

So the AI wave has done two things. It stripped the final remnants of value out of the static PDF format, and it also handed marketing teams the ability to build interactive tools without engineering support.

The tool Clay used is called Ploy, built by Bryant Chou, who co-founded Webflow in 2013 and was its CTO until last year. That's over a decade spent letting marketers build websites without waiting on an engineer. Ploy reads your existing site to learn your brand, then builds and hosts against that, so the website does work on its own where it used to sit there as an asset somebody had to maintain.

I pointed Ploy at my agency website pistachio.so a couple of weeks ago. It found my existing arguments about marketing's “missing middle” between publishing content and creating pipeline, and proposed a resource hub built around that concept. Then it went ahead and built the pillar page with a diagnostic questionnaire that scores which layer of your own system is leaking.

Clay's four steps

Alex Lindahl is Clay's Creator in Residence (find me a cooler job title, I'll wait…) and he calls Ploy their secret weapon.

Lindahl recently broke down the 4-step process they use at Clay to create and launch highly effective lead magnets. Get your notebook out, these are worth stealing.

1. Pick a use case your product already answers. Slow replies to inbound leads, mostly due to insufficient or inaccurate data, is one of the main problems Clay's product solves. So measuring slow replies is a very relevant datapoint. The study demonstrates their product while it collects the data, which is what keeps it from being a vanity exercise.

2. Generate the data at scale. Clay started with a target list of over 10,000 of their highest-value prospects. An AI agent went to each site and worked out which page held the demo form, another agent filled it in, and when a company replied by email or rang back, the timestamp logged itself. Then they scored each reply as a template, lightly personalised or fully personalised, and analysed whether the company had qualified the lead correctly. The whole dataset was landing in Ploy, so there's an interactive benchmark every company can compare themselves against.

3. Build with tools that make it fast and repeatable. Ploy had already worked out Clay's design system, so it spun up a fully functional and on-brand page with no designer in the loop. It holds the dataset, so it can reactively generate an individual version for each visitor, turning one study into more than six thousand separate answers. It hosts the thing too, so there was no infrastructure to stand up and nothing to hand over afterwards. That’s four jobs that used to need engineering help now handled by one tool, and Lindahl said it's the reason they launched this whole thing within about a week.

4. Use the distribution you already have. Clay's execs post on LinkedIn and several of their employees run newsletters, so the launch goes out through people who already have an audience. Every email address the page collects goes to Clay's sales team with that company's own numbers attached, so their follow-up conversations are based on something specific and personalised.

Where PDFs can still win

A benchmark report and a personalised tool sit at opposite ends of the funnel. An industry average is quotable, which is why RevenueHero's findings appear on other sites, including trade press. It’s top of funnel.

A breakdown of a specific company's own results is basically account-based marketing, where you pick the companies you want first and build for them specifically. The lead data from the page arrives with a named account attached and a reason to contact them. It's highly targeted, but completely automated, bottom of funnel. Hex, an analytics platform used by Figma and Anthropic, runs their version on Ploy as pages built for specific accounts, so the company is decided up front and the dedicated page is ready and waiting for them.

The mistake is asking one to do the other's job. Plenty of teams gate a benchmark report and then wait for pipeline out of it, which hands a bottom-of-funnel job to what has become a top-of-funnel asset, and that has been getting harder every year.

What the research is for

Most of us have research sitting somewhere that could easily be turned into a report. It’s whatever your product sees that nobody else sees, Clay’s sales metric is only one option. Your energy bill has been doing this for ages, with the little chart showing how your house compares to similar ones nearby. Same data the provider always had, pointed at one specific customer.

For a long time lead magnets were about finding some broad and unique insight and saying it clearly, because the finding was the scarce (and therefore valuable) part. But now the finding alone isn’t inherently useful, so the scarcity has moved to the far end of the journey. Turning that finding into something specific and meaningful for the person reading it.

So go and look at whatever research you're sitting on, and ask what a single reader's own row in it would say. Most of us never asked the question, because the honest answer required engineering support we were never going to get.

That part is over. Between a PDF download and a targeted interactive tool, you already know which version you'd hand your own email address over for.

P.S. you can try out Ploy for yourself, Brand Chemistry subscribers get 50% off your first month with code Isaac-Peiris

If you enjoyed this post or know someone who may find it useful, please share it with them and encourage them to subscribe: brandchemistry.co/p/clay-lead-magnets

Reply

Avatar

or to participate