Creating Maai

We spent this week on business models, demand and competition, sourcing, AdSense, affiliates, and dropshipping. Most of it did not apply to me, and figuring out why it did not apply taught me more than the parts that did.

The sourcing material assumes you are selling a physical object. Find a supplier, compare quotes, watch out for the shady ones, decide whether you are holding inventory or dropshipping it. My class project right now is a car detailing page, so technically I have a product, but the product is my own labor, and there is no wholesaler for that. The affiliate and AdSense material assumes traffic is the asset and someone else's product is the payload. I read all of it and kept thinking: none of these are the business I actually want to build.

The one I keep coming back to is not the class project at all. It is a side thing I have been building called Chisel. Right now it is barely anything. I have been teaching myself by making apps on Base44 and scraping pieces together, and nothing solid has come out of it yet. But it is the only idea I have that comes from somewhere real.

Here is the demand read, and it is not from a keyword tool. It is from being a returned missionary. Almost everyone I know who came home is looking for the same thing. They miss Preach My Gospel. Not the content, the structure. The way your day had a shape and the shape pointed at something. So they go looking for an app that gives that back. Strive Planner. Better calendar apps. Any of a hundred planners. And they all do the same thing: they copy the features. Goals, daily check ins, weekly review, progress tracking. Every one of those exists in ten apps.

You get the skeleton, the boxes and the streaks and the checkmarks, with no soul in it.

But not one of them connects to a central purpose. There is no personal mission statement underneath it. That is why people download these apps and quit them in three weeks. You are not maintaining the structure, because the structure is not attached to anything.

So the competition is not weak. There are a lot of calendars and they are well made. The gap is not a feature gap. It is a purpose gap, and I do not think you can close it by adding another feature.

On the model, I think freemium is where I land, at least for now. If the whole premise is that this only works when it is connected to something you actually care about, then a paywall in front of the first experience defeats the point. Someone has to get far enough in to find out if it means anything to them. What I do not know yet is what sits behind the paid line, and I am suspicious of the obvious answer, which is to lock features. Locking features is exactly the skeleton problem again, just with a price on it.

That is the honest state of things this week. The class project is a detailing page. The real project is not a project yet. But I know what is wrong with everything else in the category, which is more than I had a week ago.

↑ Close

This week was Project #1, choosing a business idea, plus picking a site builder and going through site design guidelines.

Project #1 makes you brainstorm twenty ideas, then run them through a screen. Google Ads opportunity, suppliers, startup cost, weekly time, profit potential. You score each one, weight the criteria based on your situation, and let the math tell you what to build. I weighted time and startup cost heaviest because I am a student with a job and not much money, and I needed something that could realistically produce results inside two and a half months.

The winner was a local car detailing lead generation site. Score of 82 on paid opportunity. The idea was to be the place people in Rexburg look when they need their car detailed, with a bench of independent contractors I could route jobs to based on what the customer needed. High buyer intent, low local competition, most of the ads showing up are national chains that do not know Rexburg exists. On paper it is a good idea. The spreadsheet was not wrong.

Here is what I noticed only after I finished. Of my twenty ideas, three of them were Chisel wearing different hats. Number six was a habit tracker printable shop. Number fourteen was a journaling and self growth newsletter on a subscription model. Number nineteen was a personal development resource hub. The thing I actually think about all day showed up three separate times in a list I wrote by hand, and not one of them made my top five.

They lost fairly. Journaling newsletter: hard to gain subscribers quickly. Personal development hub: requires ongoing SEO. Habit tracker: need design and promotion. Every one of those cons is true, and every one of them is a variation on the same sentence. This takes longer than two and a half months.

The screen worked exactly as designed. It selected for what can be proven inside a semester.

Car detailing can be proven inside a semester because the loop is short: someone searches, someone clicks, someone books, money moves. Chisel cannot be proven inside a semester because the whole claim is about whether a person is still using it in month four. The metric that matters most does not exist yet on day sixty.

I do not think the tool is broken. I think I need to remember that it optimized for a constraint that belongs to the class and not to me. The class ends in August. I do not.

On the site builder, I picked Wix. Comparing tools, I did not really have an argument for anything more complicated. I needed a landing page for a lead gen business, it needed to look legitimate enough that someone would trust it with their car, and I did not want to spend my limited hours fighting with a tool instead of running ads. Wix does that. I am fairly sure I let ChatGPT do most of the actual page assembly.

The part I keep chewing on is that in the same week I picked a drag and drop builder for the class project, I was hand writing HTML and CSS for Chisel. Same person, same seven days, opposite decisions. And I do not think I was being inconsistent. For the detailing site the website is a container. Nothing about it is the product, so the fastest container wins. For Chisel the interaction is the product. How the thing feels when you touch it is the entire argument. You cannot drag and drop a feeling.

That is probably the most useful thing I learned this week and it was not on the assignment. The right tool depends on whether your website is the point or the packaging.

↑ Close

This week was ecommerce platforms, payment gateways, sales tax, and legal structure.

The ecommerce half did not apply to me. Shopify, WooCommerce, Stripe, Square, connecting a cart to a gateway, figuring out how a platform handles fifty different state sales tax rates. All of it assumes money moves through your website. On a lead generation site it does not. Someone searches, lands on my page, and I hand them to a detailer. The money moves in a driveway. My site never touches it.

So I watched the material and filed it. Looking back, this is the second week in a row where the class taught me something that did not fit the thing I chose, and I did not stop to ask why that keeps happening.

The legal structure half was more interesting even though it was hypothetical. We were not required to actually form anything, just to decide what we would pick and defend it. I picked an LLC.

The textbook reason to pick an LLC here is liability, and it is a real reason. I was planning to route customers to independent contractors who would be doing physical work on other people's cars. If someone puts a scratch in a bumper, that is a question about who is responsible, and the answer should not be my personal savings account.

But that is not actually why I wanted it. I wanted an LLC because I could put more than one thing under it. I want to build more than one business. I want to have more ideas than this one, and I liked the idea of a structure that could hold all of them instead of a structure built around a single project.

The LLC appealed to me as a container for things that do not exist yet.

Which is worth sitting with, because at that exact moment I was split. I could not tell you whether I was a car detailing lead gen guy or a Chisel guy. The class had made the choice for me on paper. Inside my own head it was not decided. So of course I liked the entity that does not force you to answer. An LLC lets you say I am going to build several things without saying which one is real.

That is either good instinct or a way to avoid committing. I think it might be both. There is a version of this where wanting the flexible container is just wanting to not choose, and I know that about myself well enough to write it down.

What I notice now is that Stripe was in the material this week. A subscription business needs a payment gateway, and Chisel eventually became a subscription business. The answer to a problem I would have later was sitting on the slide in front of me and I did not see it, because I was not thinking about Chisel as a business yet. It was still a thing I did at night.

The class content and the real project were in the same room this week and did not talk to each other.

↑ Close

This week was the Idaho Business Wizard, sales tax and remote seller rules, and Project #2, which was the actual site build.

I built it and submitted it. sum25009.wixsite.com/my-site-4. Thirty out of thirty five, and the note back was website without its own domain name. That is the honest grade. I did not buy a domain, so I lost the five points for not buying a domain, and there is not much to reflect on there.

What I want to write down is the part that did not cost me points.

The site had a button that said get a quote today. The button did nothing. You could click it and nothing happened. It sat there on the page looking exactly like a button that works, in the right place, styled correctly, saying the right words, and behind it there was nothing at all.

Nobody caught it, because nobody clicked it. The grader was checking for pages, navigation, and an attempt at design, and I had all three. The rubric was satisfied. The business was not real for a single second.

That button is the whole thing, and I did not see it at the time. I built a shell of a business. Pages that looked like a business has pages, a nav bar because businesses have nav bars, a call to action because you are supposed to have a call to action. Every feature of a lead generation site was present. The purpose was missing. There were no detailers on the other end. There was no process for what happens when someone actually wants their car detailed. There was just the shape of it.

I wrote in my first post that every planner app copies the features and skips the purpose. Three weeks later I built a skeleton with a dead button on it and turned it in for a grade.

The site existed for about two weeks and then I let it go.

The Business Wizard was the same kind of empty. I ran it and nothing came of it, because for a lead gen site with no revenue and no product there is very little to register. The sales tax material had the same problem it had last week. Remote sellers, nexus, fifty states with fifty rates, all of it assuming money moves through my website. Money was never going to move through my website. Money was never going to move at all.

So here is the actual lesson learned and it is not flattering. I chose the idea that scored highest on a spreadsheet, built the minimum artifact that would satisfy a rubric, and never once did the thing that would have exposed it, which is call a single detailer in Rexburg and ask if they wanted leads. One phone call would have told me whether this was a business. I did not make it, because I did not want the answer. I wanted the grade, and I got the grade.

The class is teaching me to build businesses. I used it to build an appearance.

↑ Close

This week was responsive search ads, keyword lists, and go live with your site. I did not do most of it, because this is the week I quit the car detailing site.

The reason was not strategic. It was that the next step was to put my own money into Google Ads, and I looked at that and realized I did not want to spend a dollar on this. Not because the numbers were bad. The numbers were the best numbers I had. Opportunity score of 82, real buyer intent, national chains asleep at the wheel in a town they have never heard of. If you handed that spreadsheet to someone else they should probably run the ads.

I just did not care about it.

That is worth being precise about, because it would be easy to dress this up as a business decision and it was not one. Nothing new showed up in the data. Nothing about the market changed between the week I picked it and the week I dropped it. The only thing that changed was that the class stopped asking me to describe a business and started asking me to fund one. And when it came time to actually pay, I found out what I actually believed.

There is a version of this where quitting is just quitting. I want to leave room for that. I did not make one phone call to a detailer. I did not test the idea and find it wanting. I built a page with a dead button, got my thirty points, and bailed the moment it asked something of me. That is not a heroic story about finding your calling. Some of it is that I did not want to do the work on something I did not love.

Ad spend is the first moment in this class where belief costs money. You can write a good spreadsheet about something you are indifferent to. You cannot write a check for it.

So I switched to Chisel. Which meant I was now doing the class project on the thing that lost my own screening exercise three weeks ago, for reasons that were all correct. It takes longer than two and a half months. The metric that matters does not exist by August. All of that is still true. I chose it anyway, which means I have decided the class constraint is not the constraint I am actually optimizing for.

I did not get much of the assignment done. No twenty keyword list worth defending. No campaign. No ads live, and none for a while yet. Instead I started actually building, working with Claude Code and writing HTML directly, and for the first time in this class I was working on something at night without being told to.

I am behind now. I traded a site that existed for a site that does not, six weeks into a ten week project, and the honest tradeoff is that I gave up my head start to work on something I care about. I do not know yet whether that was courage or self indulgence. Ask me in August.

↑ Close

This week was conversion tracking with Google Tag Manager, and Project #3, which was starting the ads campaign.

I got GTM installed on the Chisel page. Container in the head, conversion action created in Google Ads, tag and trigger wired up so that a waitlist signup would report back as a conversion. It fired in preview mode. It worked.

This part was different from the Wix weeks, and it took me a minute to figure out why. On Wix you install GTM through a settings panel. On Chisel there is no settings panel. It is one HTML file I wrote, so installing GTM meant opening my own file and putting the container in my own head tag. Nothing was abstracted away from me. I understood what conversion tracking was for the first time, not because the article was better, but because I had to physically put it where it goes.

Screenshot slot · Campaign #1 keywords, all zeros

Then Project #3. I built the campaign. Eleven keywords, all broad match, and reading them back is like reading my own diary through a keyword planner. Personal growth tools. Daily reflection app. Life planning app. Mindful productivity. Organize my life. Goal planning app. Habit tracker app.

Habit tracker app. That was idea number six on my brainstorm sheet, the one that lost. Now I am bidding on it.

The RSAs went to chisel-9wh.pages.dev. Goal Setting Made Practical. Reflect On What Matters. Plan Your Best Day. Google's asset panel kept telling me what I was missing: no logo, no business name, no sitelinks, no callouts, no price asset, no location. A dozen ad formats greyed out with a warning triangle saying add these to unlock. Which is a fair description of the whole project. There is no business name because there is no business. There is no location because there is no location. Google has a slot for every part of a real company and I could fill in almost none of them.

The dashboard read all zeros. Zero clicks, zero conversions, zero dollars, because nothing was running yet.

The instrument does not produce truth. It only reports what happens. I had built a very precise way to find out something I had not found out yet.

And what it eventually told me was zero. Not zero because the tracking broke. The tracking was fine. Zero because nobody signed up. Twelve people clicked the ad and none of them wanted it, and I know that cleanly, with no ambiguity about whether the pixel fired, precisely because I did this week's homework correctly.

That is the strange gift of conversion tracking. It removes your ability to tell yourself a story about the results. If I had skipped GTM I could be sitting here right now saying maybe some people signed up and I just could not see it. I do not get to say that. I built the instrument first, so the number that came back is real.

Week six I built a button that went nowhere and nobody noticed. Week eight I built tracking that works perfectly and it told me nobody cares yet. I would rather have the second one.

↑ Close

This week was ad rank, Quality Score, and keyword optimization. The assignment was to make five changes to my keywords and be prepared to defend them in class. I did not make the five changes.

I want to start there instead of burying it, because the interesting part is not that I skipped homework. It is what I was looking at while I skipped it.

The campaign had been running about a week. Fifty four impressions, six clicks, a dollar twenty five average cost per click, seven fifty spent, zero conversions. Small numbers. But for the first time in this class the numbers were about the thing I actually built, so I read them differently than I read the detailing spreadsheet.

Screenshot slot · Search terms report

Two keywords did everything. Organize my life took five clicks and six dollars and fifty two cents of my seven fifty, at a ten percent click through rate. Free habit tracker app got one click at a hundred percent click through, which sounds impressive and means one person saw it once and clicked it. Everything else, goal planner, personal growth tools, best time management, sat at zero. Not zero conversions. Zero clicks. They never got picked up.

So basically one keyword spent my whole budget and produced nothing.

The part I keep rereading is the search terms report. This is the list of what people actually typed before Google decided to show them my ad, and it is not what I expected.

Lifeat io. Deepstash. Mind movies. Xtiles templates. Those are people searching for specific products by name. They are not looking for an app like mine. They already know what they want and they are trying to get to it, and Google put me in front of them because I bid broad match on everything and broad match will show you to anyone in the neighborhood.

Mensajes positivos. Productividad personal google. Spanish. My ad is in English and my landing page is in English and I am paying to appear for Spanish searches.

Then this cluster. Daily affirmation for today. Daily reflections for today. Positive quotes for today. How to be a better person. A balance life.

Somebody sat down and typed how to be a better person into a search bar. That is not a person shopping for software. That is a person having a moment. And my ad said Goal Setting Made Practical.

That is the gap and I did not fully understand it until I saw the two lists side by side. My headlines are pitched at somebody evaluating a productivity tool. The people my ad reached were not evaluating anything. Half were looking for a competitor by name, some were in the wrong language, and the ones with real intent were reaching for something and my ad was a category label. Nobody clicks a category label at 1am.

Zero conversions on six clicks is not enough data to prove anything. I know that. But the search terms are not a sample size problem. That is a list of real sentences from real people and it tells me who Google thinks I am.

So why did I not make the changes. The fixes are obvious enough that I can write them down right now. Kill broad match. Add Spanish negatives. Add competitor names as negatives. Pause the keywords with zero impressions. Rewrite headlines to sound like the daily reflection searches instead of like a productivity brochure.

I think the reason is that I was busy building. I spent this week in the HTML, not in the ads dashboard, and I told myself the product needed to be right before the traffic mattered. There is a real argument in that. There is also a real chance it is the same instinct as week seven, where I would rather work on the part I enjoy than the part that tells me something uncomfortable.

The data was sitting there the whole week saying you are talking to the wrong people in the wrong language. I read it. I did not act on it. That is the honest record of week nine.

↑ Close

This week was responsive search ad optimization and the RSA project, which is fifty points and the biggest ads assignment of the class.

I did the work. Four screenshots, every question answered, evidence pulled from the actual campaign. Reading it back it is the best analysis I have produced in this class, and I mean that as a real observation and not a compliment to myself, because of what happens at the end.

Here is what the report says. Match types are mostly wrong. Almost every keyword is broad match on a one dollar a day budget, which is the worst possible combination, wide net and no money. Mind movies pulled three clicks and three dollars and thirty six cents, the most expensive term in the whole campaign, for a manifestation product that has nothing to do with what I built. Mensajes positivos took a dollar ninety four in the wrong language. Habit tracker is actively misleading, because I removed habit tracking from the product in favor of a needs based model, so anyone searching that term lands on a page that does not do the thing they searched for. I am paying to disappoint people.

Screenshot slot · View asset details, headline performance

The headline data was the part I did not expect. Chisel Personal Growth App carried seventy four impressions and six clicks. Half of every click in the entire campaign came through the one headline that just says the name of the app and what it is. Build Better Daily Habits got three. And then the ones I liked best, Reflect On What Matters and Live More Intentionally, got two or three impressions each and zero clicks.

Join The Chisel Waitlist got twenty six impressions and zero clicks. Twenty six people saw me ask them to join a waitlist and not one of them did.

The vague poetic headlines lost. The plain one won. People do not click a philosophy from a search results page.

I have been building something whose entire pitch is that it is about purpose and not features, and the headline that performed was the one that named the product and its category like a normal ad.

Then the last question, what conversion rate would make this campaign work, and this is where the math got honest. Sixteen dollars spent, twelve clicks, a dollar thirty three a click. If a waitlist signup is worth two dollars to me, I need roughly a sixty seven percent conversion rate to break even. Landing pages convert between two and ten percent. So at my current cost per click and my current assumed value, this campaign is not close to viable, and it is not close by a factor of about ten.

There are only two levers. Get the click cheaper, or make the signup worth more. And a free waitlist signup is not worth much, which is the quiet thing underneath this whole project.

So I had a full diagnosis with the prescription written next to it. Exclude mind movies. Exclude the Spanish terms. Pause habit tracker. Switch broad match to phrase match. Split the ad group. Rewrite the headlines toward the plain ones that actually worked.

I added the negative keywords. That is all I did.

Not the match types. Not pausing habit tracker. Not the ad group split. Not the headline rewrite. I turned in a fifty point document that says exactly what is wrong with my campaign, and then I fixed the easiest item on the list and left the rest sitting there.

Last week I did not make the changes because I had not really looked. This week I looked, wrote it all down in detail, got graded on it, and still mostly did not act. That is worse, and I would rather write it down than pretend the assignment counts as the work.

I think I know why. Fixing the campaign optimizes for something I do not actually want. If I switch everything to exact match and rewrite the headlines to say Chisel Personal Growth App, I will get slightly cheaper clicks to a waitlist for a product that twenty six people already declined to join. The ceiling on that is a handful more signups on a page nobody wants. The report told me the campaign becomes possible when the value of a signup and the cost of a click move close enough together. I cannot move the click much. So the real work is on the other side of that equation, and it is not in the ads dashboard. It is in the product.

That might be true. It is also exactly what someone would say to avoid doing a chore. Both can be right at once.

↑ Close

This week was the AI traffic strategy assignment. Take everything the RSA campaign taught you, interview an AI about it, have the AI critique its own recommendation, then pick a new way to get traffic and run it for three weeks.

I did the whole brief. Business summary, RSA summary, demand diagnosis, three strategies considered, measurement plan, risks and assumptions. And then in section E, Selected Strategy, the place where you name the thing you are actually going to do, I wrote to confirm in class.

Cost: to confirm in class. Time investment: to confirm in class. Business model: confirm pricing structure in class if this has been decided.

The whole document is built around a hole where the decision goes.

Some of that is real. The assignment says your instructor may require approval before you begin, and the brief was due before that conversation. But I know myself well enough to see the pattern. Week five I wanted an LLC because it let me hold several projects without picking one. Week eleven I wrote a strategy brief that names three good options and commits to none of them. Same move.

The analysis underneath it is not bad. The demand diagnosis is the part I actually believe. My earlier read was that this is an awareness problem, that nobody knows a thing like this exists, so I have to create demand. The RSA data says no. People are searching goal planner and daily reflection app right now. The demand exists. Twelve of them clicked. Zero signed up.

So the bottleneck is not traffic. It is that people arrived and did not want it. More traffic to a page nobody converts on is just paying to be rejected faster, and the honest strategy is to fix the page and the offer before buying another click.

Which is roughly what I did, except not as a strategy. On June twenty seventh I started rebuilding the whole thing from scratch and renamed it Maai. New HTML, built to feel more like the app I actually miss, the one from my mission. Not a planner with reflection bolted on. The thing itself.

I did not frame it as the answer to the assignment. But it is the same answer. The campaign was not failing because of match types. Twenty six people saw Join The Chisel Waitlist and none of them joined. That is not a keyword problem. That is the product not being worth joining yet.

Then the part that is not flattering. I did not really implement anything this week. The assignment wants three weeks of measured activity and I was buried in my public speaking class and stressed, so the traffic experiment mostly sat there. I rebuilt the product instead, which is the thing I wanted to do, and let the thing I was assigned to do wait.

There is a version of me that says this is prioritization, that I identified the real bottleneck and went after it. There is another version that says I was stressed and did the fun part. Looking at the brief with three blanks in the middle of it, I do not think I get to claim the first one cleanly.

The strategy document does say one true thing though, and it is the thing I keep circling. The campaign becomes viable when the value of a signup and the cost of a click move close enough together. I cannot move the click. So everything depends on the signup being worth something, and it will not be worth anything until the thing behind it is worth wanting.

That was worth rebuilding for. It just was not what was due.

↑ Close

This week was the two week action plan and the message and creative clinic. Bring one real asset. Not a plan for an asset. A thing that exists.

The strategy I landed on is to interview returned missionaries on camera and put it on Instagram. Not an ad for the app. Just the problem, said out loud by someone who has it. You come home, the structure is gone, the days blur, and you go looking for something that gives it back. Then at the end, do you feel like him, go look at maai.pages.dev.

I think this is the first genuinely right idea I have had in this class, and I want to be careful about why, because I have been wrong in a confident sounding way before.

Every other thing I have tried started with the product and went looking for people. Keywords are the product looking for people. Headlines are the product looking for people. Even the good headline, Chisel Personal Growth App, is the product introducing itself to a stranger and hoping the stranger is interested. That is why twenty six people saw Join The Chisel Waitlist and nobody joined. They had no reason to. I was asking them to care about my thing before I had shown them I understood theirs.

The interviews go the other direction. Nobody watching has to be convinced the problem is real. They already know.

The app is just the thing at the end of a sentence they were already nodding at.

It also fixes something the ads could never fix. My whole first post was about how every planner copies the features and skips the purpose. Skeleton, no soul. But an ad headline is thirty characters. You cannot put purpose in thirty characters. It comes out as Live More Intentionally, and that got two impressions and zero clicks, because as a headline it is a greeting card. A person telling you what it felt like to come home and lose the shape of their day is not a greeting card. The channel finally fits the thing I am trying to say.

And it is aimed at a real, concentrated audience instead of everyone on the internet who typed a productivity word. My search terms report was full of Spanish speakers and people looking for manifestation videos. That is what buying broad reach gets you. Returned missionaries are a specific group, I know exactly where they are, and I am one of them.

So the diagnosis is right, the channel is right, the audience is right, and the message is right.

I have not shot a single interview.

That is week twelve. The plan is good and the plan is all there is. The clinic asked for one real asset and the join page exists, which is something, but the strategy itself is still entirely a document.

I notice this is now three weeks in a row of the same shape. Week nine, I saw the data and did not act. Week ten, I wrote the full diagnosis and only did the easy part. Week eleven, I wrote a strategy brief with the decision left blank. Week twelve, I found the right strategy and did not start it. My thinking has gotten sharper every single week and my execution has not moved.

I do not think that is a motivation problem. I want this one. I think it is that everything I have not done requires another person. Call a detailer. Point a camera at a friend and ask him to talk about coming home. All the work I actually finished this semester, the HTML, the analysis, the spreadsheets, the documents, is work I can do alone at night. Everything I stalled on has a human being on the other end of it.

The class taught me to find demand, price a click, read a search terms report, write a hypothesis. Useful. But the thing standing between me and the first honest test of this idea is not knowledge. It is asking somebody for twenty minutes.

Ten weeks of posts and the lesson learned is that I know what to do and I keep not doing the part that requires courage. Writing that down is not the same as fixing it. But the interviews are not hard, they are just uncomfortable, and I am out of analysis to hide behind.

↑ Close