How I Built k8mak.com

Most personal sites start with a template and end with a landing page nobody visits. I wanted something different: a site that demonstrates how I think, not just what I have done. So I built k8mak.com the same way I build products for clients -- with real strategy work, not aesthetic preferences dressed up as decisions.

This post walks through the process: what I did, what I built, what I cut, and what I learned about applying product discipline to your own work.

The process

I used a 15-step product strategy process that I have refined across dozens of client engagements. The same sequence I would run for a $2.7B CRO client or a seed-stage health tech startup, I ran for myself.

The steps: JTBD and persona research, market sizing, competitive analysis, north star metric definition, an OKR workshop, opportunity solution trees, RICE scoring, roadmap construction, journey mapping, a pre-mortem, security review, PRD writing, story mapping, and a coherence audit. Fifteen steps, each producing an artifact that feeds the next.

Not every step was equally useful for a personal site. But four shaped nearly every decision that followed.

Jobs-to-be-done research

I identified seven jobs this site needed to serve. Two were obvious: career transitioners looking for AI tools, and independent consultants building their practice infrastructure. But the most important job emerged late in the process -- a VP or Director of Product evaluating whether I am the right hire. I named her Hannah.

Hannah does not want a skill catalog or a free tool. She wants to see how I think, how I have handled complexity, and whether my experience pattern-matches to her open role. That single persona reshaped the entire site architecture. Portfolio case studies moved from marketing afterthought to the highest-priority content on the site.

The other revelation: after leaving my last role, I realized I was not just serving Marcus (the solo consultant persona). I was Marcus. Going independent validated every assumption I had made about that persona's pain points, because I was living them. "I built this because I needed it" is the strongest positioning a product person can have.

OKR workshop

I ran a proper OKR workshop against a 90-day window. The objective: establish k8mak.com as a credible professional surface and validate one paid product.

Four key results:

1. Receive 5+ inbound inquiries (interview requests or consulting leads) attributable to the site
2. Publish 6+ blog posts demonstrating product thinking, with at least 2 ranking page 1
3. Launch one product (a Consulting Starter Pack at $49-79) and generate $500+ in revenue
4. Achieve 100+ monthly qualified engagements by day 90

The north star metric -- Monthly Qualified Engagements -- was the most important decision. It counts portfolio deep reads, tool completions, blog engagement, contact form submissions, and purchases in a single number. One metric that captures whether the site is doing its job across both audiences.

RICE scoring

I scored every feature candidate and the results surprised me. The "Why Breadth" framing section -- three paragraphs on the About page explaining why my cross-domain career is a strength, not a lack of focus -- scored a RICE of 5,000. Higher than almost every feature that would take ten times longer to build.

This blog post scored 5,760. A portfolio narrative polish scored 1,600. The "How I Work" page scored 1,714.

The pattern: high-impact, low-effort content work consistently outscored new features. RICE does not care about what is fun to build. It cares about what moves the metric.

Pre-mortem

Before writing a single line of code, I ran a pre-mortem. I imagined it was August 17 and the project had underperformed. What went wrong?

The risks I identified:

The "Why Pay?" problem. Career transitioners can get 80% of what my tools do from ChatGPT. The free tier needs to be good enough to build trust, and the paid product needs to deliver something genuinely harder to replicate.

Solo operator overload. I had 8-12 hours per week for this work, down from the 20 hours I estimated in April. That constraint meant cutting scope aggressively -- one paid product instead of three, four sprints instead of six.

Stream B might not work. There is no guarantee hiring managers visit candidate websites. If they do not, the entire portfolio-first strategy is wasted effort. The mitigation: UTM-tagged resume links so I can measure whether anyone actually clicks through.

The dual-stream tension. Is a site that serves both job seekers and tool users actually two separate strategies wearing a trenchcoat? I decided the overlap was genuine -- the same blog post that impresses Hannah also ranks for search terms that reach Priya -- but I flagged it as the biggest strategic risk.

What I built

The site ships as a Next.js application on Vercel with markdown-driven content. The stack is deliberately boring: Tailwind for styling, gray-matter for frontmatter parsing, Vercel Analytics for tracking. No CMS, no database, no authentication layer for the public site.

What is on it:

Interactive tools. An AI Maturity Assessment, an AI Health Check, a Get Unstuck coaching tool. These serve the career transitioner and consultant personas. They also demonstrate that I build things, not just talk about building things.

Portfolio with case studies. Outcome-first case studies across 14 years of engagements. Each one structured as Situation, Approach, Impact -- the format hiring managers scan when deciding whether to call.

Blog. Writing on product leadership, AI strategy, team coaching, and the work itself. Dual-purpose: thought leadership for Hannah, SEO for organic discovery.

Services and engagement models. Three ways to work together, designed around outcomes rather than hours billed.

A 340+ skill AI agent system. The site runs on Skillet, my consulting practice skill library. Every tool, every content generation workflow, every audit template is a reusable skill that I also use with clients. The site is not a brochure for my work -- it is an artifact of my work.

What I cut

RICE scoring and the pre-mortem killed more features than they approved. That is the point.

A testimonial carousel. Low-impact, high-maintenance. Client quotes are more credible embedded in case studies where they have context.

A pricing calculator. Sounded clever in brainstorming, scored poorly in RICE. Prospects who need a calculator are not ready to buy.

Three paid products. The original plan included a PM Case Study Kit, an AI Playbook, and the Consulting Starter Pack. The revised plan ships one. Validate the model before scaling the catalog.

A newsletter signup. Deferred, not killed. Email nurture requires Resend integration and ongoing content commitments. At 8 hours a week, I cannot sustain both a blog cadence and a newsletter. The blog serves the same audience-building purpose with lower operational cost.

An interactive resume timeline. The resume is a PDF. Hiring managers expect a PDF. Overengineering the format does not make the content more compelling.

Every cut traced back to the same logic: does this move a key result, and can I sustain it at 8-12 hours per week? If the answer to either question was no, it was out.

The dual-stream model

The most unusual structural decision was running two strategies in parallel on the same site.

Stream A is the product business: free tools that demonstrate capability, a paid Consulting Starter Pack, and eventually a broader digital product catalog. This stream targets career transitioners and independent consultants.

Stream B is the professional surface: portfolio case studies, blog posts that show how I think, and enough social proof that a hiring manager can evaluate me without a phone screen. This stream targets VPs and Directors making hiring decisions.

The streams share infrastructure. The same blog post that ranks for "product strategy framework" also demonstrates strategic thinking to Hannah. The same portfolio page that shows Priya my engagement history shows a recruiter my industry breadth. The same interactive tools that serve Marcus also prove to Hannah that I build, not just advise.

This is not an accident. The OKR workshop surfaced that both streams serve the same north star metric: Monthly Qualified Engagements. A portfolio deep read from a hiring manager counts the same as a tool completion from a career transitioner. The site succeeds when people engage deeply, regardless of which door they came through.

What shipped and how fast

The entire product strategy process -- all 15 steps, all artifacts -- ran in roughly two weeks of part-time work. The site build followed across 6 sprints. At the time of writing, the site has 20+ published blog posts, 5 interactive tools, a full portfolio, and a services section.

AI tooling was a force multiplier. I used Claude to run every step of the product strategy process, to draft and refine content, to build the site itself, and to write this post. The skill library that powers the site is also the skill library I used to build it. That is not a parlor trick -- it is a proof point. If I am going to sell AI-powered consulting tools, the most credible thing I can do is use them myself, visibly.

The lesson: shipping velocity matters more than polish. A site with real content and working tools outperforms a beautiful site with placeholder text. Every week the portfolio sat unpublished was a week Hannah could not find me.

What is next

The foundation is down. The next 90 days are about validation: do hiring managers actually visit candidate sites? Do career transitioners complete the tools? Does the Consulting Starter Pack sell?

If you are a product leader thinking about building your own site, my advice is simple: run the same process you would run for a client. Your site is a product. Treat it like one. Define the jobs it needs to do, score the features, cut what does not move the metric, and ship before you are comfortable.

If you want to see the work, browse the portfolio. If you want to try the tools, they are free. If you want to talk about what you are building, get in touch.