The second post in this series was about plumbing: hostnames, canonical tags, fonts, images. This one is about the shape of the content itself, and it starts with a count I should have run months ago.
FlowHaxa is built on four stages: Control, Optimize, Automate, Build. Every guide lives in the hub for one of them. So I counted the guides in each hub.
The map
Before: guides by stage Control 6 ███ Optimize 29 ██████████████ Automate 13 ██████ Build 5 ██ Plus 12 short answers. Control: budgeting Optimize: cash flow Automate: tools Build: planning
The middle of the framework held more than twice as many guides as both ends combined. Control, where a reader is supposed to start, had 6. Build, where they're supposed to end up, had 5.
How a reasonable plan tipped over
No single decision did this. The Optimize hub started empty. Then it got a 15-article batch, because cash flow is the site's core topic. Then it kept growing, because cash-flow questions have the most search demand, and every new article there felt like the safe bet. Each choice made sense on its own. Together they built a site that's strongest exactly where the framework says a reader is halfway through.
Someone who follows the framework from the beginning finds a handful of budgeting guides, a crowded middle, and a thin ending. The journey runs out of road right where it should pay off.
What else the audit found
- A quarter of the pages were thin. One page in four had fewer than 300 words. Most were the 12 short answers, at around 120 words of actual answer each, plus the two original calculators with almost no text around them.
- The newsletter led nowhere. The sidebar on every article said "Cash Flow Hacks, Weekly" with a Subscribe button. The button went to a page that didn't exist.
- A guide had gone out of date. The free accounting software guide said Wave's free plan included receipt scanning. Wave's own pricing page now puts receipt scanning on a paid add-on and automatic bank imports on the paid Pro plan.
- No analytics. There was nothing to measure behavior with, so every decision so far had been made on page counts and search data alone.
What changed
The ends got filled in
Six new guides, three for each thin stage, each with a worked example and links into the calculators:
- Control: how to make your first budget, a step-by-step zero-based budget, and a subscription audit.
- Build: short- and long-term cash flow goals, planning cash flow around major life events, and saving for a down payment without wrecking the monthly budget.
After: guides by stage Control 9 ████ Optimize 30 ███████████████ Automate 13 ██████ Build 8 ████
That's better, not balanced. Optimize is still the heaviest stage, and I'm fine with that: it's where the demand is. The goal was never equal numbers. It was making sure a reader who starts at the beginning, or comes looking for the ending, finds enough to keep going.
How the new guides were chosen
I didn't have to guess at topics for the Build stage. The planning hub's configuration already listed five guides in a Life Events and Goals group that had never been written. The hub quietly skipped them, so nobody noticed the gap. Three of the new guides fill those slots. The budgeting guides came from the other end: they answer the questions a reader has before any of the existing Control guides make sense, starting with how to make a first budget at all.
Every new guide also follows the rule from the first post: link forward to the next stage. The first-budget guide points to the zero-based method and the calculators. The goals guide sends readers to the emergency fund calculator before it gets to longer-term goals.
The short answers got longer
Each of the 12 short answers kept its direct answer at the top and gained a worked example, steps or a checklist, and a link to the matching calculator. They went from about 120 words to roughly 290 to 380. The two thin calculator pages got an explanation, a worked example and a FAQ.
The site got measured
| Measure | Before | After |
|---|---|---|
| Pages under 300 words | 1 in 4 | 7% |
| Words published | 42,059 | 55,245 |
| Median page length | 559 words | 604 words |
| Calculators | 2 | 5 |
The newsletter button goes somewhere now
There's a real /newsletter page, and every article ends with a short sign-up form. Addresses are saved to a file one folder above the public website, so nobody can download the list by guessing a URL. A hidden field catches bots, and each visitor is limited to five sign-ups an hour.
Analytics, and a way to see what earns
GA4 went on every page with this update. The tag records the visit immediately but loads Google's script after the page finishes, so it doesn't undo the speed work from the last post.
FlowHaxa also got its first recommendation boxes: 21 pages, from the accounting and payment guides to all five calculators, now show the tools each guide actually discusses. Every click is recorded as an event in GA4, with the product and the page.
None of those boxes earns anything yet. I haven't joined the affiliate programs, so each button simply goes to the company's site. I built it that way on purpose. When a program approves me, I paste one link into one file, and that button picks up the sponsored label and the disclosure automatically. Until then, the click data tells me which boxes deserve a program at all.
The Wave guide was fixed before a box went on it. A recommendation sitting next to an out-of-date claim is worse than no recommendation.
What I'm not claiming
I don't have results to report yet. These changes are days old, analytics started with them, and Search Console needs weeks to show whether rebalanced hubs get more impressions. Anything I said about traffic right now would be a guess.
A note on scoring: In my own portfolio scoring, which weighs market, content, technical health, monetization, moat, timing and risk, FlowHaxa moved from 64.6 to 72.5 and from sixth to fourth of eleven sites over these changes. That score measures readiness, not revenue. It says the site is in better shape to earn, not that it does.
What I'm measuring next
- Hub engagement. Do readers who land in Control or Build keep going to the next stage?
- Recommendation clicks by page. Which boxes get clicked, and which get ignored and should come down?
- Newsletter sign-ups by source. Every sign-up records the article it came from.
- Search visibility by hub. If the new Control and Build guides start collecting impressions, the rebalance worked. If they don't, demand really is concentrated in the middle, and I'll weigh the ends differently.
What's still undone
The four verbs still aren't named on the site. Readers see four category names, not a sequence, which was the gap I pointed out in the first post. Build is still the smallest stage. Both are next.
What I'd tell someone running a framework-based site
Count your content by stage, not by category, and do it before you write the next batch. Demand will always pull you toward one part of the journey. That's fine, as long as you notice and keep the rest of the road open. And check that every button you show a reader leads somewhere. A newsletter that goes nowhere costs you subscribers. It also quietly tells readers the site isn't finished.
The same approach shaped the HSAMaxGuide rebuild: stop adding pages, look at the network, then fill the gaps the map shows. The FlowHaxa project page has the current numbers, and SEO site architecture covers the method.