The Factory Floor
What we're building, testing and learning.
Notes from active projects across web, AI, search, product and immersive work.
Why AI Forgets Things It Can Still See
Why ChatGPT and Claude lose track of decisions that are still in the conversation: context windows vs. attention, lost in the middle, stale versions, compaction and RAG, and a five-step workflow for long AI projects.
Read the post-
AI Automation for Small Business: What to Automate First
Six small business workflows to automate first, with the tools, monthly subscription costs, the point where a person reviews the output, and how to measure hours saved.
-
OpenAI Dots Review: Our First Day, Plus What the Docs and Early Testers Show
We used OpenAI Dots on real agency work for a day, then checked OpenAI’s docs and early testing. How Dots work, where they broke, what to try first.
-
OpenAI Dots: What Always-On AI Agents Mean for Business
OpenAI Dots are always-on AI agents with their own cloud computer and 4,000+ app connections. What they do, who gets them, and where custom agents still fit.
-
Agency Pricing Models in 2026: What Replaces the Hour
Five agency pricing models compared: hourly, retainer, fixed fee, productized subscription, and outcome-based. What replaces the billable hour, and the 90-day switch.
-
The Agency Business Model in 2026: Deliver or Compound
The agency business model is shifting from deliver-and-move-on to accumulate-and-compound. What actually compounds, how to measure it, and why most agencies have not.
-
Agency Utilization Rate in 2026: Formula, Benchmarks, Limits
Utilization is billable hours over available hours, and the 2025 average hit a record-low 66.4 percent. The formula, worked examples, and why AI breaks the metric.
-
Agency Operations in 2026: Build a Production System
Agency operations in 2026 is a four-layer production system: research, planning, execution, and QA, with people placed where the risk concentrates. How we built ours.
-
Marketing Agency Profit Margins in 2026: The Real Numbers
Agency net margins averaged 13 percent in 2025, down from a 15 percent norm. The benchmarks by headcount, why AI efficiency cut margins first, and where the recovery is.
-
Local SEO Checklist: What to Do First and Why
The 26-item local SEO checklist we run with clients, in order: profile fields, page signals, citations, reviews, links, and the numbers worth tracking.
-
AI-Native SDLC: The Agency Reality Check
Anthropic published two pieces this year that together describe what happens to an engineering organization once agentic coding stops being a tool somebody uses…
-
Ox Alpha: 1M Tokens, No Name, and Real Risk
Ox Alpha shipped August 20 with a 1,048,576-token context window and no named developer. Here are the numbers and what they mean for client code.
-
Style LoRA Training: What It Actually Takes
A client came to us with a library of flat vector illustrations they’d built up over several years. The work was consistent: same line…
What's next should be worth building.
New products. Better systems. More compelling experiences. Whatever the opportunity, we bring the thinking, creativity, and technical expertise to make it real.
Start a Conversation 4520 SE Belmont St, Suite 300Portland, Oregon