Episode 012: If Nobody Knows It Exists

In the last episode, I wrote about Operations and the point where making something once has to become making it repeatedly. A design can work perfectly on the bench and still fail as a business if every build depends on remembering exactly what happened the last time. Materials have to be available, parts have to be organized, packaging has to work, quality has to stay consistent, and the whole process has to survive a normal workday instead of only working when everything goes right. That was a major step for JW Designs because it moved the business from simply having products to having something that could actually be produced.

Unfortunately, figuring out how to make the product consistently did absolutely nothing to make people know it existed. I could design a better lid, make it reliably, package it properly, put it on the website, and still have almost nobody see it. That became the next problem. It was also one of the places where AI and the Business OS started becoming much more useful than I originally expected, not because AI could magically market the product, but because it could help me understand all the different ways people were discovering, ignoring, reacting to, or completely missing what I was building.

Making It Was the Part I Understood

JW Designs started with a product problem, and product problems make sense to me. I had enclosures that needed better lids, I could see ways to improve them, and I was comfortable working through the physical issues involved. Design the part. Print it. Cut the acrylic. Test the fit. Change something. Test it again. Engineering gives you feedback you can usually see or measure. Operations gives you another kind of feedback. Can I make the next one the same way? Did the packaging survive shipping? Did the process take too long? Is there a better way to organize the parts? Even when something goes wrong, the problem is usually sitting somewhere in front of me.

Marketing did not feel nearly as natural, and it was one of the first places where AI looked like it might offer an easy shortcut. Ask AI to write a Facebook post. Ask it for a product description. Ask it for an SEO title. Ask it for an email. Within seconds you can have something that looks like marketing. The problem was that generating words was never really the difficult part. The hard part was figuring out what was worth saying, where the right customers were, what language they were using, which conversations mattered, what needed follow-up, and whether any of the effort was actually producing useful attention.

That is where my use of AI started changing. Early on, it was easy to treat it like a content generator. Over time, that became one of the less interesting things it could do. The Business OS gave all of the other information somewhere to go. Customer comments, search terms, show conversations, product reactions, social posts, retailer discussions, SEO changes, and failed experiments could start becoming parts of the same operating picture instead of isolated moments I was trying to remember.

Marketing Was Bigger Than Posting

At first, it was easy to think about marketing mostly in terms of social media. Take some pictures, make a post, put something on Facebook or Instagram, and hope somebody sees it. Social media is part of the job, but it turned out to be a very small definition of what Sales and Marketing actually needed to do. A customer searching Google for a replacement lid for a specific enclosure is a marketing opportunity. Choosing a product name that uses the same words that customer is likely to search is another. Organizing the website around manufacturers and enclosure families matters. Product photography matters. Search indexing matters. Showing up at a reptile show with physical products matters. Talking to another keeper matters. Somebody else using the product and mentioning it to a friend matters.

Even the words on a product page matter because being discovered is not very useful if the customer arrives and still cannot figure out whether the lid fits the enclosure sitting in front of them. We spent time cleaning up product names, reorganizing collections, working on compatibility language, improving page titles and descriptions, and trying to think about the website from the customer's point of view instead of mine. Internally, a SKU or engineering name might be perfectly clear. That does not mean anything to somebody typing the name of their enclosure into Google.

The Business OS made that work easier to think about because Sales and Marketing could become more than a place where I asked for the next post. It could watch the business for attention opportunities. A recurring search phrase could point toward an SEO problem. A question from a customer could expose a weak product description. A reaction at a show could reveal language worth using elsewhere. A conversation with another business might actually belong in Business Development. Something somebody noticed about a product might need to go all the way back to Engineering. The value was not simply producing more marketing. It was connecting what we were learning.

That eventually became the way I started describing Sales and Marketing inside the Business OS: an Opportunity Engine for Attention. The department is not there just to keep a posting calendar full. Its job is to keep looking for reasonable ways the right people might discover that JW Designs exists.

The Boring Stuff Worked Better

For all the different things we have tried, the best return on the effort so far has probably come from one of the least exciting parts of marketing: SEO. Not a flashy campaign, not an influencer program, and not some clever social strategy. Mostly a lot of tedious work making sure the website describes the products in a way that people are actually likely to search for them.

That meant changing product naming, organizing the site around enclosure manufacturers and families, cleaning up page titles and descriptions, improving compatibility wording, submitting pages for indexing, and paying more attention to the phrases people might use when they were frustrated with the lid already sitting on their enclosure. It was not glamorous work, but it has been the clearest example so far of effort connecting with people who were already close to the problem we solve.

That makes sense in hindsight. JW Designs sells a very specific product. Most people will never need one of our lids, and getting in front of a large general audience is not automatically valuable. Search is different. Somebody typing the name of their enclosure along with words like replacement lid, acrylic lid, humidity lid, or something similar is already telling us something important. They are looking for a solution. We do not have to convince them that they should care about enclosure lids. They already care. We need to make sure JW Designs can be found when they go looking.

AI has been useful here because it can help examine naming, customer language, product structure, and the way the website is organized. The Business OS gives that work continuity. Instead of doing SEO once, changing some page titles, and forgetting why we made the decisions, Sales and Marketing can preserve what we learned and build on it. That turns SEO from a one-time cleanup into an ongoing feedback loop.

The Audience Matters More Than the Size

SEO has not been the only thing that has worked. Some of the better results have also come from highly relevant communities where the audience is already close to the problem. Tarantula Talk has been a good example. A post there is not going in front of a random audience. It is going in front of people who understand enclosures, ventilation, humidity, visibility, and all of the tradeoffs that come with traditional screen lids. The product needs less explanation because the people seeing it already understand why somebody might want an alternative.

That has reinforced another lesson: relevance has mattered more than raw reach. A relatively small group of people who actually own the kind of enclosure we make a lid for can be far more useful than a much larger audience with no reason to care. Search works for the same reason. Shows work in a different way because the right people can physically handle the product. Word of mouth works because the recommendation comes from somebody other than me. These are all different forms of attention, but they are useful because the audience is already close to the problem.

That does not mean we stop looking for creators, influencers, retailers, or other ways to expand the audience. We are still looking for people who make sense. We still put products in hands at shows. We still place samples when the relationship and audience seem right. We still watch for communities where the product fits naturally. What has changed is that we do not have to pretend every one of those channels is equally proven.

SEO has stronger evidence behind it right now. Relevant communities have shown value. Shows and product-in-hand exposure have produced strong reactions and useful relationships. Creator and influencer work is still something we want to develop, but the results so far have been much more uneven.

The Tools Were There. The Customers Were Not Necessarily There.

One of the deceptive things about running an online business is how many tools are available to help you sell. Shopify has an entire ecosystem built around this. There are retailer connections, affiliate and influencer tools, discount codes, analytics, sales channels, product feeds, integrations, and all kinds of ways to connect another seller or another audience to your products. On paper, some of those tools looked almost perfect for JW Designs.

We set up Shopify Collective with another business. The plumbing works. Products can be listed by another retailer without them holding the inventory, the retailer gets a defined margin, and JW Designs can still fulfill the order. For a small manufacturer, that sounds like a very efficient way to expand distribution. So far, it has not turned into a meaningful source of sales. We have also looked at influencer and affiliate programs. Again, the idea makes sense. Give someone with the right audience a reason to share the product, give their audience a discount, give the creator a commission, and let Shopify handle the tracking. Reasonable idea. Technically possible. Not especially fruitful yet.

That has been another useful lesson for both AI and the Business OS. It is very easy to confuse a configured tool with a working business channel. An account can be connected, the products can be available, the margin can be defined, and the entire technical setup can be complete without a single customer showing up because of it. AI can help research the platform, compare the economics, draft the outreach, structure the program, and track what happens. It cannot provide the audience on the other side.

That does not mean I am throwing those tools away. Shopify Collective could still make a good retailer relationship easier to operate. An affiliate system could make a real advocate easier to compensate. The infrastructure may become very useful when the right relationship exists. What I do not want to do is mistake that infrastructure for proof that a marketing strategy worked.

Some Opportunities Go Nowhere

This is also where the story gets less tidy, and I think it needs to stay that way. I do not want Building a Better Business to become a polished retrospective where every experiment somehow turns into part of a brilliant plan. That is not what building this business has looked like. Some things work better than expected. Some produce almost nothing. Some opportunities look promising right up until they quietly disappear.

We have put products into people's hands who genuinely liked them and still watched the expected exposure never happen. In one case, a sample placement looked like it was going to lead to useful visibility on a live selling platform. The people involved liked the lid. The reaction to the product itself was strong. The opportunity seemed legitimate. Then they simply never got the chance to feature it. Nobody did anything wrong. Nothing dramatic happened. It just did not turn into what we thought it might. That particular marketing opportunity was basically a bust.

There have been other versions of the same lesson. Conversations that seemed promising and never developed. People who expressed interest and never followed up. Posts that barely moved. Channels that looked like they should matter more than they actually did. None of that means the original experiment was necessarily stupid. Sometimes the idea was reasonable and the result was still nothing.

The Business OS helps here in a way that is less exciting but probably more valuable than a lot of AI marketing promises. It helps us remember that those things were misses. It gives us a place to preserve what actually happened, change the status of an opportunity, compare one experiment with another, and stop rewriting history six months later as though everything worked. AI can help analyze the result, but it cannot change the result.

AI Cannot Make Somebody Care

This may be the most important boundary I have found in using AI for Sales and Marketing. AI can help identify an opportunity. It can help organize the information around it. It can help research language, structure a product page, draft a message, compare ideas, preserve relationship history, and remind the system that something still needs attention. It cannot make somebody care.

It cannot force a creator to feature the product. It cannot make a retailer follow up. It cannot make a customer search for exactly the right phrase. It cannot make an algorithm put a post in front of the right person, and it cannot create trust simply because it generated a polished message.

The reptile shows have been a good example of that boundary. Putting an EncloSure lid directly in somebody's hands does something no product description can do. They can feel the weight, work the latch, look through the acrylic, and see how the enclosure changes when the original screen lid is replaced. Some of those conversations have produced relationships. Some have produced useful feedback. Some may eventually turn into something larger. Some were simply good conversations.

AI helps me capture and understand what happened afterward. The Business OS helps make sure the useful parts of those conversations do not disappear into a weekend I barely remember six months later. Neither one replaces the actual human interaction that made the conversation valuable in the first place.

That is why the lesson from SEO working well is not to stop doing everything else. We still look for influencers. We still take products to shows. We still place samples. We still pay attention to retailer opportunities. We still participate in relevant communities. The difference is that the Business OS gives us a way to separate what is proven, what is promising, and what is still an experiment.

The Business OS Made the Mess Manageable

A one-person business generates a ridiculous amount of information. One weekend at a couple of shows can produce product reactions, retailer conversations, potential partnerships, sample placements, pricing feedback, engineering questions, names to remember, people to follow up with, and several things that seem incredibly important in the moment. Then Monday comes. There are orders to make, messages to answer, products to design, materials to buy, a website to maintain, and in my case a full-time job waiting for me too.

That is the kind of problem the Business OS was being built to solve. AI could help unpack a weekend afterward, separate the useful signals from the noise, identify who actually needed follow-up, distinguish a marketing contact from an engineering issue, and preserve enough context that the next conversation did not have to start from scratch. It did not mean AI suddenly knew which opportunities would succeed. It meant I no longer had to rely entirely on my own memory to keep every opportunity alive.

That difference matters because AI on its own can actually create more work. There is always another post it can write, another campaign it can suggest, another idea it can generate, and another thing I could spend time doing. That is the opposite of what I need. I already have more possible work than available time. A useful system should help reduce that problem, not make it worse.

The Business OS gives AI boundaries and context. Sales and Marketing can look for opportunities, preserve evidence, compare results, and make recommendations. I still decide what is worth pursuing. That is the balance I am looking for. AI helps me see more of the business. The Business OS keeps that information organized. Real-world results decide what deserves more attention.

What Has Actually Worked So Far

At this point, I do not think JW Designs has discovered a magic marketing formula, and I would be suspicious if I thought we had. What we have is evidence. SEO has produced the clearest bang for the effort so far. Relevant communities such as Tarantula Talk have been fruitful because the audience is already closely matched to the problem. Shows and direct product exposure have generated strong reactions, useful relationships, and real-world feedback, even though the conversion from those interactions is harder to predict. Word of mouth has value because somebody else is providing the trust. Influencer outreach and sample placement still make sense, but the results have been inconsistent. Shopify Collective and affiliate-style infrastructure are available and may eventually matter, but they have not produced much yet. Broad social media helps maintain visibility, but it has not been the primary engine driving the business.

That list will probably change. In fact, I hope it does. A creator relationship could suddenly work. A retailer could turn Collective into something meaningful. A show contact could become a major partnership. A new community could become a strong source of customers. The point is not to declare one channel the winner forever. The point is to know what the evidence says today while continuing to run reasonable experiments.

That is what I want the Opportunity Engine for Attention to do. Not manufacture noise. Not chase every platform. Not fill a calendar because the calendar exists. It should keep looking for places where the right people might discover JW Designs, while paying enough attention to the results that we learn where our effort is actually worth spending.

Then Somebody Actually Notices

Eventually, somebody does notice. A customer places an order. A store owner asks a question. Another business sees something that might fit what they are doing. Somebody recommends the product to somebody else. A conversation that started as marketing begins turning into something that looks more like a relationship.

That is where the problem changes again.

Sales and Marketing can create the introduction. AI can help preserve what happened before that moment, recover the history, compare the opportunity with everything else going on, prepare for the next conversation, and make sure a good lead does not quietly disappear because I got busy doing something else. The Business OS can keep the opportunity connected to the rest of the business.

But once another person or company starts seriously discussing how we might work together, we are no longer dealing only with attention. We are dealing with trust, economics, expectations, commitments, and relationships. That is a different kind of work, and it became the reason JW Designs needed another department.

Next episode: Business Development, and what happens when attention starts turning into actual opportunities.