Last Week

Last week, I paused restaurant visits long enough to reconsider how I was finding potential Shokken customers.

The first dozen conversations had divided restaurants into a few broad groups. Some already had a waitlist system and no reason to replace it. Some did not have enough regular demand to justify adding another tool. Others only expected the problem to become important during a busier season. The conversations were useful, but they also showed that a list of restaurants with complaints about long waits was not a strong enough targeting strategy.

I also used the pause to experiment with a small group-ordering web app and a goal-driven development process. That work reinforced the value of small, testable slices and independent review, but Shokken’s immediate challenge is no longer building more product. It is getting the existing product in front of the right operators.

This week, I broadened the marketing surface by researching a more targeted and more scalable way to reach them.

From Door-to-Door Visits to Better Timing

Shokken is published on both Android and iOS, and it is available for restaurants to download and test. My current goal is to onboard the first operators who are not simply volunteering to help me test it.

So far, I have taken a white-glove approach: visit restaurants in person, speak with an owner or front-of-house manager, and demonstrate the app when there is interest. That produces detailed feedback, but it is difficult to scale. Even a short conversation can take fifteen to thirty minutes, while a full demonstration and the questions afterward can keep me at one restaurant for close to an hour.

Las Vegas summer weather adds a very practical constraint. Walking between restaurants in roughly 110-degree heat makes each trip physically draining, and repeated cold introductions take an emotional toll too. In-person outreach may make more sense later in the year, when the weather is more comfortable and seasonal restaurant traffic begins to increase. It cannot be my only path in the meantime.

The better opportunity may be to approach restaurants when they are already changing. A new restaurant, a new location, or a business going through an ownership or management transition has fewer entrenched processes and more reason to evaluate its operating tools. The goal is not to contact every restaurant. It is to identify the smaller group that may actually need a walk-in waitlist and reach them at a moment when trying one is reasonable.

What does it mean in English?

Walking into random restaurants is a little like knocking on every door in a neighborhood because one house might need what I am offering. It works eventually, but most of the conversations begin without enough context.

Public business records can provide a better starting point. A newly registered business name does not tell me whether the business is a full-service restaurant, a counter-service shop, a cafe, or something else. But I can combine that record with the business’s website, social media, and other public information to build a clearer picture.

That picture lets me answer the questions that matter before I interrupt anyone:

  • Is this actually a restaurant with walk-in traffic?
  • Is it likely to have a waitlist problem?
  • Is it new or going through a transition?
  • Is there a useful way to contact the owner or manager?

Only after a business looks like a plausible fit would I prepare a tailored email and potentially follow up by phone. This is still cold outreach, but it is different from sending the same message to every food-and-beverage business in a database. The research should make the message more relevant and avoid wasting time for both sides.

Nerdy Details

Public data becomes useful when the pieces are joined

No single public record contains the complete lead I need. One source may include a registered or fictitious business name and an email address. Another may contain business details. A website or social account may explain what kind of restaurant it is and whether its service model could benefit from Shokken.

The useful result comes from joining those pieces together. In database terms, that sounds straightforward. In practice, names are inconsistent. One file may include punctuation that another omits. The registered name may differ from the name displayed to customers. Spacing, abbreviations, and formatting can all prevent an exact spreadsheet lookup from finding a match.

The data also changes. A one-time manual spreadsheet could answer a small question, but repeating the same cleanup and matching work every time a source is updated would quickly become tedious. At the other extreme, building a full lead-generation platform would be far beyond what Shokken needs.

The right scope is a small internal workflow: ingest a limited set of public records, identify likely matches across sources, evaluate whether each business is a plausible Shokken customer, and prepare useful contact information and a personalized draft message. It does not need to become a general-purpose product or a service that sells contact data.

Exact matching is not enough

Some steps are deterministic. A parser can normalize whitespace, punctuation, or capitalization, and ordinary code can join records that share a stable identifier.

Other steps involve judgment. Two slightly different names may refer to the same restaurant, but they may also refer to unrelated businesses. A website may describe dining in vague terms. A business category may be technically correct while revealing very little about whether guests wait for tables.

That is where a reasoning model may help: not as an unquestioned source of facts, but as a way to evaluate ambiguous matches and summarize the evidence already gathered from public sources. Any output still needs to stay tied to that evidence. A polished guess is not a qualified lead.

I have not built this workflow yet. This week was research: identifying possible data sources, understanding how the records fit together, and reducing the idea to a plan that is small enough to implement.

Social media is another surface, not an excuse to rebuild everything

A viewer also suggested maintaining a regular stream of social media videos. I agree that Shokken needs a broader online presence, but producing good short-form video is its own discipline. The strongest product videos have deliberate pacing, framing, music, sound effects, and editing. I cannot assume that publishing clips consistently will be effortless just because I can write software.

I previously built a programmable video pipeline using Remotion, Playwright, and mobile automation. It could capture the app and assemble a specific short demonstration without requiring me to edit every frame manually in DaVinci Resolve.

The problem is that I designed the pipeline too tightly around that one video. Producing a substantially different video would require changing so much of the system that manual editing might be faster. Generalizing it into a reusable video-production framework is possible, but that would become another large engineering project.

That is the trap I need to recognize. Building systems is comfortable for me; selling is not. If I am not careful, I can keep inventing infrastructure projects that feel productive while postponing the conversations the business actually needs.

Next Week

Next week, I plan to implement the smallest useful version of the lead-research workflow. The immediate goal is to ingest selected public data, qualify potential restaurant matches with supporting public information, and produce a short list of contacts for tailored email and phone outreach.

Social video still belongs in the broader marketing plan, and in-person visits may become more effective again as the season changes. But neither requires a new platform before I take the next step. For now, the priority is simple: spend less time building ways to market and more time actually putting Shokken in front of restaurants that may need it.