Hiring an Engineer for an Old House? Ask Who Has to Build From the Report
Finding outside expertise for an old building has changed in ten years, and so has the way a plausible report can look better than a good one.
The letter arrives as a PDF, five pages, with a cover photograph of the front elevation and a stamp on the last sheet. It reads well. It cost less than the two other quotes, arrived four days sooner, and answers the question you asked, which was whether the sagging floor in the back room is a problem. What it does not do is tell the mason what to build, or tell the building department what was assumed about the foundation nobody excavated. That gap is where most of the money gets lost, and it is invisible at the moment you pay the invoice.
What the search looked like ten years ago, and what it looks like now
A decade back, finding a structural engineer or a preservation consultant for a house built before 1940 usually ran through a contractor, a local architect, or the person at the historical commission who had seen forty of these. You got a name, sometimes two. The engineer came out, spent an hour in the crawlspace, and sent two pages on letterhead that referred to the actual room by the actual name you used for it. The report was thin because the visit was thick. You could not compare five providers because you could not find five, and the recommendation carried the reputation of whoever made it.
Now the search starts in a browser and returns more names than you can process. License lookup is a genuine improvement: nearly every state board publishes an online register, so verifying that a professional engineer's stamp is current and unencumbered takes about four minutes rather than a phone call and a wait. Reports are longer, better illustrated, and often include drone elevations or a laser scan of a floor plane. Remote and hybrid reviews became normal after 2020 and never fully reversed. Presentation got cheap. Judgment did not.
The practical consequence is that a plausible report and a good report now look far more alike on the first read than they did in 2015. Both are formatted. Both have photographs. Both cite a code edition. The difference has moved out of the appearance of the document and into a few specific places inside it: what the site visit actually covered, what was assumed rather than observed, and whether the recommendations are written in a form somebody can price. The Federal Trade Commission oversees how services are advertised to consumers, which covers the claims on the website, but not whether the engineering underneath is any use to you.
The party almost nobody considers when they hire
Most people choose a consultant as though the report is the end of the transaction. It is not. The report is an input to somebody else's work, and the person who has to build from it is the party missing from the hiring decision entirely. That is usually a contractor, sometimes a mason or a timber framer, occasionally a plan reviewer at the building department who will decide whether the permit issues. None of them were in the room when you picked. All of them will spend your money interpreting what was written.
Ask a contractor to price a vague report and you get one of two things: a high number that carries the ambiguity as contingency, or a low number that will be revised the week the wall opens up. Ask three contractors to price a specific report and the spread between them narrows, because they are all bidding the same work. That narrowing is the return on a good engineer, and it is measurable in a way that the quality of prose is not. If you want a single test before you hire, it is this: send a redacted sample report from the candidate to a contractor you trust and ask whether they could bid it without calling anyone.
What to ask before you sign the proposal
Ask who performs the site visit and who signs the report, because on larger firms these are frequently different people and the gap between them is where local knowledge disappears. Ask how long the visit will be and what will be opened, moved, or probed. Ask what the report will assume rather than verify, and get the answer in writing, since a report that quietly assumes a competent footing under a 1912 rubble foundation is not wrong so much as untested. Ask whether one round of follow-up questions from your contractor is included, and what a second visit costs.
Then ask the question that separates the two categories cleanly: will the deliverable include a repair scope specific enough to bid, or is it an opinion of condition? Both are legitimate products at different prices. Confusion between them is the single most common reason a homeowner pays twice, once for the assessment and again for somebody to turn it into instructions. Say which one you are buying in the proposal itself.
Reading the deliverable when it lands
Look first at the limitations section, which is usually near the back and usually skipped. It tells you what the professional declined to stand behind, and in an old building that list is where the real risk lives: concealed framing, prior repairs of unknown date, moisture behind plaster, anything under the slab. A short limitations section on a complicated house is a warning, not a reassurance. A long, specific one paired with a recommendation for targeted opening up is somebody telling you the truth about what a visual inspection can establish.
Then check that every recommendation names a location, a member, a dimension or a performance requirement, and a sequence. A report that says the rear bearing wall needs support is an observation. A report that says which joists, over what span, with what temporary shoring, before which demolition step, is a document your contractor can build from and your inspector can approve.
Old buildings reward the consultant who has seen the same failure in the same era of construction thirty times, and that experience shows up in the specificity of the scope rather than the polish of the cover page. Hire for the sentence your mason will read on a Tuesday morning, and the rest of the choice tends to make itself.