Spring Hill KS, 66083
A customer calls and asks why nobody answered her. She filled out your contact form twice. You check the inbox, and there's nothing from her. The site looks fine. It has looked fine the whole time.
That phone call is what I build for. A website isn't a launch. It has to keep working when nobody is watching, and the first thing to break is usually the form. I do this for businesses across the Kansas City metro, from Spring Hill to Overland Park.
In August 2026, I sent a test message through each contact form on a nonprofit's site. The screen said "sent" every time. Not one message arrived. There was no bounce and nothing in the spam folder. The site's own server was sending the mail as if it came from a Gmail address. Gmail throws that kind of mail away without a word. A third form, one I had built three days earlier, delivered every test. That's how I knew the site was fine and the two older forms were not. Nothing on the screen could have told you.
The other shape of the same Tuesday: in 2018, I set up a site and handed it to the host. It went live, and within the hour the owner wrote to say three large orders had already come in. Before that launch email went out, someone had sent a test through the form and watched it reach a person. That one test is the whole difference between the two stories.
So before any form I build goes live, I send a real test message and watch it land in someone's inbox. I do it again after launch. That's why this page exists.
The record · August 2026
A nonprofit's site, August 2026 · what its own contact forms did
of the test messages I sent. Both forms told me “sent” every time, and I found no bounce and nothing in spam.
said “sent” on every submission
form, built three days earlier, delivered every test
between the form that worked and the two that did not
Source: the site's mail server and Gmail's own rejection rule, read 2026-08-26. The screen said sent. The inbox said nothing.
Two of my clients learned this on sites nobody was touching.
In June 2026, I opened the weekly crawl report for a longtime client and found 435 error pages. The week before, there were 24. Nobody had changed anything on purpose. Something had updated in the background, and the site rebuilt itself around it.
The record · June to August 2026
A legacy client's site, June to August 2026 · what the weekly crawl alerts counted
in one week. The week before, the crawl report showed 24, and nobody had touched the site since.
with no alt text
with no headline
made on purpose by anyone
Source: the crawler's weekly alert mail, three threads, read 2026. Something updated, and the site rebuilt itself around it.
Then on August 23, a retailer's Search Console showed a new warning: "Blocked by robots.txt." Google had stopped indexing some of its pages. The site was getting about 700 clicks a month from Google. Nobody had touched it. Its plugins had been updating themselves every night.
A site nobody checks drifts away from the one you launched. That's why a monthly check is part of the work, not an add-on.
Before you hire anyone
Three times a new site is the wrong spend, and what to do first instead.
If it loads, the form delivers and the copy is yours, a redesign is the expensive way to fix a small thing.
A designer cannot fix an empty site. I audited a shop whose blog had comment spam showing and not one photo of their work.
Take 30 photos of your work and write one page of plain text about what you do.
The problem is what happens after the form: who gets the email, and how fast. A new site will not fix that.
The honest answer for all three is to spend the money somewhere else first. Ask me and I will say so.Robert Dove, Dove Web Consulting
Ask meA site on a platform you can log into, with the pages your customers use, built to load fast on a phone. Kansas City businesses mostly need five to ten pages done properly, not fifty.
Before any design, I ask you for thirty photos of your real work and one page about what you do, in your own words. Selling online? Picking the platform is its own decision: Creating a modern eCommerce website.
Last August, a nonprofit's forms said "sent" and delivered nothing. So on every site I build, I fill out the contact form myself and check the inbox, before launch and again after.
Analytics and Search Console go on before launch day. If they go on a month late, that first month of visits is never recorded.
The domain, the hosting, and the site login in your name, written down, handed over on launch day. Nothing about your site should depend on remembering my email address.
Error pages, updates, form delivery and index status, read once a month and reported in a page you can read in a minute. The two records above are what this catches.
At a plumbing company I worked for, finished jobs sat open in the dispatch software. They stayed open until somebody remembered to close them. Nobody was assigned to watch for that. Most businesses have a few jobs like it, and today they run on memory and good intentions.
The fix is a small computer in your own cloud account, in your name. It runs those jobs on a schedule and tells you when one breaks. At that plumbing company, it closed finished jobs every half hour, so nobody had to remember. It also carried the ad click from the website form into the dispatch record, so the office could see which ad brought each call.
The record · 2026
Three clients' own cloud servers, 2026 · a plumbing company, a bathroom remodeler, a leadership-development firm
that must match byte for byte on every deploy: the local file, the server file, and what the public gets
scheduled on the plumbing company's server; 21 of 22 scripts healthy at the write-up
between runs of the remodeler's check: 27 things that must stay true, each proved by breaking it
a month: the budget alert the firm set on her own cloud account, about what the plumber's server costs
Source: each client's own server records and deploy rules, read 2026. Every server sits in the client's own cloud account, in the client's name.
After that, the jobs run whether anyone remembers or not, and you get one page a month saying what ran and what broke. The plumbing company's server cost about fifty dollars a month. I quote the build after I read what has to run. One rule came from a mistake. A document was reported live for two days while the server still held the old copy. The check had only read the status code. Now every deploy is checked three ways, the local file, the server file, and what the public actually receives, byte for byte.
Do you need a web designer? Three questions.
1. Do you have a page of plain text about what you do, and photos of your own work?
2. Does your contact form reach a person within a business day, and has anyone tested that this year?
3. Do you hold the logins for the domain, the hosting and the site itself?
Nothing you click here is sent anywhere. It runs on this page only.
Three questions, twenty seconds. If the answer is no, this page just saved you a build.
Tune-up $300 to $1,500, refresh $1,000 to $3,500, rebuild $2,000 to $4,500. Which one you need is a decision, and the post that walks it through is Do you really need a new website?. The number on this page is the build. The monthly check is priced on the pricing page.
For a five- to ten-page site, plan on eight to twelve weeks once your photos and text are in hand. I can't start the clock before that, because nobody else can write what your business does.
The one you can log into and update yourself. I build on WordPress and MotoCMS, and I've taken over sites on Wix and OpenCart. What matters more is that the login is yours and somebody checks the site after updates run.
Yes, and it's in writing: the domain, the hosting, and the site login are in your name on launch day. Ownership is a bigger subject, so it has its own article on this site.
Assessment, planning, build, test, launch. The phases are written out in What is the most effective web design process?. This page is about what happens after the last one.
Contact Us
Other Services