Check the Buyer: finding what stops construction companies getting in.

A LinkedIn outreach experiment at Qonstrue, which provides on-demand expertise and managed services for construction and physical industries. I owned the targeting, prospect sourcing, message testing and measurement, and reported the results weekly.

Do construction sales teams respond to help finding tenders they can already bid for, or to help working out what a specific buyer requires before they can get approved, listed, invited or specified?

The second promise, if it held, would widen a company's addressable market rather than just improve its tender search.

Four earlier message versions (takeoff automation, free tender assessment, lost-bid review, and access and eligibility) produced 93 connection requests and 10 accepted (11%) over a month. Moving the promise to access, and the audience from bid and estimating roles to business development and sales roles, lifted acceptance to 21% in a week. That became Check the Buyer.

• Sourced and enriched the audience through ZoomInfo and Clay: 586 contacts added.
• Ran the outreach on LinkedIn through GetSales: connection requests with no note, and one message only after a request was accepted.
• Tested the sales-team audience against a directors and founders variant and a smaller-business variant, all pointing to the same Check the Buyer offer.
• Tracked every stage from request to reply to pipeline, week on week.

• Reach307 connection requests, 113 accepted (36.8%).
• Engagement143 messages sent, 76 opened (53.1%).
• Response18 replies (12.6% of messages sent), and 6 opportunities reached the pipeline stage.
• Best audienceThe sales-team variant: 302 requests, 107 accepted (35.4%), 72 messages opened and 14 replies. The directors and founders variant had only 5 requests, too few to compare.

Acceptance rose from 11% across the first versions to 36.8%. Replies and pipeline entries are early signals of demand, not signed customers.

• The promise and the audience moved the numbers, more than wording did. Sales teams responded to access and eligibility far more than to tender search.
• Opens and replies held up at volume, which a one-week sample could not show.
• The most useful number to track is connection requests sent, not leads loaded into the tool.

Check the Buyer became the main outbound offer, with a finished landing page and LinkedIn copy inside the platform's word limits. A related experiment, opportunity readiness, tests what could stop a company before it commits a team to a specific job. I also set up a plain-language weekly report so the wider team could read results without a data background.

I built Qonstrue's internal AI tool for go-to-market, a GTM operating system on Claude and Supabase. It connects research, company discovery, audience qualification, outreach and performance data. I also document it. I also run the weekly product and website analytics report.