Six Weeks to the First Order: What a Real Website Build Includes, and the Four Hours I Need From You

Most people call me because they need a website.

That is the right instinct, and it is also where a lot of money gets wasted, because a website on its own does not make you findable, does not make your email arrive, and does not let anyone give you money. The site is the thing you are buying. Making it work in the real world takes about forty more steps that nobody itemizes for you, and most shops will happily build the site and leave those forty to you.

So here is the whole thing. The build itself, and everything I set up around it, written out for a company that could be any supply house in the country. If you want to run it yourself, run it. I would rather you have the map than not.

The company

Kestrel Supply Co. is made up. I built it as a composite of the kind of business I work with most, because using a real client's name to sell my own services is tacky and I am not doing it.

Kestrel is a second-generation plumbing and HVAC supply house. One warehouse, a counter, eleven employees, about $2.8 million a year, essentially all of it over the phone or across that counter. Dana runs it now. Her father ran it before her. She is very good at her business and has no interest in becoming good at any of what follows.

What Dana wants: contractors should be able to order online instead of calling, and people searching for a part at 7am should find Kestrel instead of the big-box store forty minutes away.

Day one is a phone call, and mostly I listen

The first conversation is ninety minutes and I am not selling anything during it. I want to know how money actually enters this business, because that determines what gets built.

With Kestrel, the answer reorganized the whole project. Roughly 70% of revenue comes from about forty contractor accounts who buy on account, pay within thirty days, and already know exactly what they want. The other 30% is walk-in retail. Those are two completely different products sharing one website. The contractors need fast reordering and their own pricing. The walk-ins need to find the place and know if the part is on the shelf.

If I had skipped that question and built what Dana asked for on the first call, she would have gotten a handsome catalog that her forty best customers would never use.

I also ask the unglamorous questions. Who owns the web address. Who set up the current email. Is there an existing Google listing and can anyone still log into it. What does "in stock" mean here, and does the number in the computer match the number on the shelf.

That last one matters more than anything technical. If your inventory count is wrong, putting it online turns a private problem into a public one.

Week one is a plan, and part of the plan is what we are not doing

You get a written plan before anyone touches a keyboard. It says what gets built, in what order, what it costs, and when it will be done. It also says what I talked you out of.

For Kestrel, the plan cut two things Dana came in wanting. She wanted the full catalog online, all nine thousand items. We launched with six hundred, the ones that actually move, because nine thousand items with messy information is worse than six hundred done properly, and cleaning up nine thousand records would have added two months and a lot of money to reach the same revenue. She also wanted a phone app, because a competitor has one. Nobody uses the competitor's app.

This is the part people are actually paying me for. Not the typing. The twenty-five years of watching which of these decisions costs somebody real money.

Here are the five that come up most.

Your web address is locked inside the website company. You bought it during signup and it lives on their account. When you want to leave, and eventually you always want to leave, you spend two weeks on support tickets or you lose the address your customers know you by.

Somebody else owns your Google listing. A marketing company set it up under their own account years ago. It is the single most common thing I find, and getting it back is a Google support process that can take weeks. That listing is often the first thing a customer sees and the thing that sends you the most calls.

The ad account lives in someone's personal Facebook profile. Usually a former employee's. When they go, the advertising history goes with them, and that history is exactly what makes your ads get cheaper and better targeted over time. It does not transfer.

Nobody ever proved your email is really yours. There are three settings that tell Gmail and Outlook that mail from your address is genuinely from you. Without them, your order confirmations quietly land in spam, customers tell you the website is broken, and the website is fine.

Nobody wrote down who owns what. This is the one that quietly costs the most. I have watched a business lose access to its own Instagram account because the only person with the password was a contractor nobody could find.

Weeks two through five: building the thing

This is the actual product, and it is where most of the hours go.

The structure follows the money. Kestrel's site has two front doors. Contractors log in and land on their reorder list, their pricing, and their open invoices. Everyone else lands on search and local information. Same website, two completely different first screens. That decision came straight out of the day-one call and it shaped everything after it.

Cleaning up product information is most of the work, and nobody warns you. Six hundred products meant six hundred rows of part numbers, barcodes, categories, specifications, and shipping weights, plus photos. Kestrel's existing records spelled the same manufacturer's name four different ways. I fixed all of that before loading it in, because loading it first and fixing it later means fixing it in three places forever. This is unglamorous, it is a real chunk of the schedule, and skipping it is why so many small-business stores look sloppy when they show up in search results.

Search that works the way people actually type. A contractor types a part number, half a part number, or the wrong name for the right part. If your search only finds exact matches, you have a catalog nobody can use. This matters more than the homepage design and almost nobody budgets for it.

Contractor pricing behind a login. Account customers see their negotiated prices. Walk-in customers see list prices. Nobody sees anybody else's numbers, and that is enforced on our end rather than just hidden on the page, because anything merely hidden on a page can be found by anyone curious enough to look.

Inventory that cannot lie. The website checks against the warehouse system so it cannot sell something that is not there. Overselling a contractor once can cost you the account.

Checkout tested with real money. We run a live card through it and then refund it. Account customers route to an invoice instead of a card payment.

Speed, because of where people actually use it. A contractor is standing in a mechanical room with one bar of signal. If the site takes eight seconds, he calls the big-box store instead. Photos get sized properly, pages stay light, and I test on a deliberately slowed connection rather than on my own fast one.

Built for the phone first. Most of Kestrel's visitors are on a phone in a truck. The desktop version is the afterthought, not the other way around.

Security as a baseline, not an add-on. The padlock in the address bar, so information travels privately. Customer data scrambled where it is stored, so a stolen copy is worthless. Spam filtering on every form, because the junk is relentless. A second step beyond the password to get into the admin. And staff logins that only reach the part of the system that person needs, so one stolen password does not open the whole building.

An admin Dana can actually use. She can add a product, change a price, update hours, and post a promotion without calling me. If the owner has to pay somebody to change a phone number, the build failed.

Checked on real devices. Test orders placed and refunded. Forms submitted and confirmed received. Every page opened on an actual phone. And the whole site run through a screen reader, the software blind customers use to browse, so they can shop too.

You own it. The code, the accounts, the data. Every asset transfers to you at the end.

And then everything around it, which is the part you did not have to ask for

A website nobody can find, whose email goes to spam, is a brochure. So all of this comes with the build. It is not an upsell and it is not a separate phase. It is what I think building a website means.

You do not need to understand all of it. That is sort of the point. But it does need to exist, and if nobody does it, it does not happen.

The foundation. Your web address moved to an account the business owns, with the transfer lock on and automatic renewal turned on, so it cannot be taken and cannot accidentally expire. Real email on your own name, so it reads dana@kestrelsupply.com instead of a Gmail address, because a generic address costs you quotes with larger customers whether or not anyone says so. Then the three records that prove your mail is really yours, the ones from the pitfall above (SPF, DKIM, and DMARC, if you want to look them up). We start those in report-only mode and watch for a couple of weeks before enforcing, because switching them on hard from day one can block your own mail. A separate administrator account from Dana's everyday email. Two-step login everywhere, with the backup codes printed and in the safe. And a company-owned password vault, not a list in my head and not a note in Dana's phone.

Proving the business is real, over and over, to five different companies. Google Business Profile is the panel that appears beside search results with your hours, phone number, and directions. It gets claimed, filled in properly, and then verified by video: one continuous take showing the street, your sign, and the counter, with location services turned on. It gets rejected for dumb reasons and you reshoot.

Search Console and Analytics show what people typed to find you and what they did next. Both go in before launch, because neither can tell you about last month if it was not running last month.

Merchant Center is what puts your products in Google's shopping results with a photo and a price. Its product list pulls straight from the store, so nobody has to remember to update a spreadsheet.

Then the Facebook and Instagram side. A Business Manager account owned by the company rather than by a person. Domain verification, which proves to Facebook that you own your own website, and which you have to finish before you are allowed to control which customer actions Facebook may track (that setting is called Aggregated Event Measurement). Then the tracking itself wired up two ways at once, through the visitor's browser and through our own server, so ad blockers and privacy settings do not silently erase half your data. Then I confirm it is matching real customers instead of assuming it works.

The places people actually look. Bing Webmaster Tools, which copies most of its setup from Google and takes five minutes. Apple Business Connect, which feeds Apple Maps and which almost nobody bothers with, so it is free ground. Yelp claimed with the categories corrected. A Nextdoor business page, which for a local supply house pulls better than Facebook does. And the boring one that beat all of it for Kestrel: manufacturer dealer locator pages. Three of Kestrel's suppliers publish lists of authorized dealers, and Kestrel was missing from two of them. That is the most motivated traffic in the entire business and it was free.

Somewhere in week four, not much happens. Google is reviewing the products. Google is reviewing the verification video. The internet is still catching up to the new address settings. The payment company is running its background check on the business. I warn clients about this week in advance, because a quiet week feels like being ignored if nobody told you it was coming.

Making it not need a person

A system that depends on somebody remembering something will eventually fail, because everybody is busy.

Kestrel's inventory updates itself on a schedule. Order confirmations and receipts send automatically from the verified address, so they land in the inbox. Review requests go out on a delay after the order is filled, not when it ships. The product list for Google refreshes itself. The site is watched around the clock, so I find out the contact form broke instead of Dana finding out from an angry customer three weeks later. Backups run, and more to the point, I restore one to confirm the backups are real. A scheduled backup nobody has ever restored is not a backup, it is a belief.

And a plain weekly email with the numbers, so Dana does not have to log into four different websites to find out how the week went.

Week six: the handoff, and the four hours

Total time Kestrel's owner spent on this project:

Ninety minutes on the first call. Thirty minutes approving the plan. Twenty minutes confirming the legal business details. Forty-five minutes walking the warehouse with a phone for the verification video. Thirty minutes across two rounds of reviewing the site. Fifteen minutes signing into things while I was on the phone with her. Twenty minutes on the handoff walkthrough.

That is a little over four hours, spread across six weeks, to get a working store and everything behind it.

The handoff is a recorded walkthrough plus a written list of every account, what it does, who owns it, and how to get into it. That document is the deliverable people underrate. It is the thing that means you are never held hostage, including by me.

The five things I am not allowed to do for you

I do as much as I legally can, and there is a hard line. Payment companies require the owner's own identity check, with your Social Security or tax number and your bank details, and you should be suspicious of anybody who offers to handle that part for you. Bank connections are yours. Anything signed as an officer of the company is yours. The verification video has to be filmed at your location, so either you film it or I drive out. And your personal logins stay personal, which is exactly why everything gets built on company-owned accounts with me added as a user, so that removing me later takes thirty seconds and breaks nothing.

Two doors at the end

When the plan is met, we sit down again, and there are two honest options.

I show you how to run it and I leave. This is the default and I mean it. The walkthrough, the list of accounts, and a system simple enough that your office manager can change a price without calling anybody.

Or I stay on, month to month, and handle updates, monitoring, and the slow improvement work. That is a line item in every quote I send, not a surprise after launch.

Plenty of clients take door one. That is a good outcome. Software you cannot operate without paying someone forever is not an asset, it is a subscription with extra steps.

If you only do one thing

Open a notebook. List every account your business depends on. Web address, email, website, Google, Facebook, the payment company, every directory. Next to each one, write down who owns the login.

Most people cannot finish that list. The gaps are where your business is exposed, and finding them costs you an afternoon instead of an emergency.

If you would rather not spend that afternoon, or you need the site built in the first place, tell me what you are working with and I will tell you what it would take. Straight answer, real timeline, no dream-selling.

Get a quote or see what I build.