Riverway Logistics
Riverway Logistics
A carrier site that answers what brokers and drivers actually call to ask
A flatbed carrier in Spring, Texas had no front door on the internet. Brokers vetting the company and drivers looking for a seat both ended up on a government lookup page. I built one that talks to both of them in about eight seconds.
The job
Riverway runs flatbed out of Spring, Texas. Steel, lumber, machinery, the kind of freight that rides on a deck instead of inside a box. Two very different people go looking for a carrier like that online, and both of them were coming up empty.
The first is a freight broker with a load to cover. Before they hand it to anyone, they want the MC number, the DOT number, the equipment, and whether the lanes make sense. They want it in about eight seconds, because they are working a list.
The second is a driver. A driver wants to know what the freight looks like, whether dispatch is direct, and who to call. Nothing else on the page matters to them.
Design that serves both means the page can’t be a brochure. It has to be a lookup table wearing a nicer shirt.
What I found
The facts existed, they just lived in four places. The MC and DOT numbers were on a federal SAFER page that looks like it was built in 2003. The lanes lived in someone’s head. The phone number was on a business card. Anything a broker wanted to confirm, they had to assemble themselves, and the ones who couldn’t be bothered just called the next carrier.
There was also a credibility problem that has nothing to do with design. In trucking, a carrier with no web presence reads as either brand new or not serious. Neither one gets you the load.
What broke
The driver application form worked on the first try, which should have been the tell.
Applications were submitting fine and the confirmation emails were going out fine. They were just landing in spam, every time, for anyone whose mail provider bothered to check. The domain was sending mail from a form handler that had no authorization to send as that domain: no SPF record listing it, no DKIM signature, nothing for a receiving server to verify against. Gmail did exactly what it is supposed to do and filed it under junk.
Nothing in the application logs looked wrong. The form said it sent. The provider said it delivered. The only way to see the problem was to look at the raw headers on a message that made it through and read the authentication results line, which said spf=neutral and no DKIM at all.
That is the failure that turned into the Google Workspace work. You cannot fix deliverability from inside the form.
The fix
The site itself is static. Every route is prerendered and served from Vercel’s edge cache, so a broker on a phone in a truck stop parking lot gets the page immediately rather than waiting on a server to think about it. The whole document is about eight kilobytes of HTML.
The structure follows the two audiences instead of fighting them:
- Verification, above the fold. MC, DOT, home base, and equipment type sit in the first screen as plain text, not inside an image, so they are copyable and indexable.
- Lanes as geography, not marketing. Texas, Gulf South, Atlantic Southeast. Named regions a broker can match against a load, rather than a claim about coverage.
- Recruiting on its own track. Drivers get a separate path with the things drivers ask about, so the broker page doesn’t get cluttered with hiring copy and the hiring page doesn’t get cluttered with compliance numbers.
- The phone number is a link everywhere it appears. Most of this traffic is on a phone. Making someone copy a number by hand is a way to lose them.
Once the domain was sending authenticated mail, the form notifications started landing in the inbox instead of spam, and the applications that had been quietly rotting in a junk folder became visible.
What I’d do differently
I’d check mail authentication before building the form, not after. The form was the fun part so I built it first, and then spent longer than I want to admit debugging an application pipeline that was never broken. Ten minutes with dig at the start would have saved all of it.
I’d also push harder on structured data. The site has the carrier’s identifiers in readable text, but a LocalBusiness block with the DOT number as an identifier would give search engines something unambiguous to attach to the name. That is a small change with a long tail, and it is on the list.