Journal
only what we
actually know ↘
AI search in 2026: why Google answers are no longer the only SEO target.
Over the past 12 months, 18–34% of our clients' organic traffic has started coming from ChatGPT, Perplexity, and Gemini citations — not clicks. We've realised that the classic SEO approach no longer answers the fundamental question: will a person actually reach your website at all?
Open articleLatest articles.
Black Friday playbook for Shopify stores — 14 sprints over 90 days.
Peak season isn't about 'figure out the creative the day before'. Our annual 90-day plan with specific deadlines, deliverables, and KPI measurement moments.
Six Klaviyo flows that generate 40% of e-commerce revenue.
Welcome, browse abandonment, cart recovery, post-purchase, win-back, VIP. And why the seventh ("birthday") usually doesn't perform the way you'd hope.
How to read GA4 attribution without losing your mind — or your budget.
Default channel grouping, custom models, and why data-driven attribution sometimes shows absurd things. An exercise with real dashboard screenshots.
Local SEO for small businesses in Riga — a complete GBP optimisation guide.
Google Business Profile audit, citation cleanup, review strategy, and what happens when you respond to negative reviews badly.
TikTok creative testing — why 90% of your UGC won't work.
Rhythmic structure, a hook patent inventory, and what happens when you confuse brand ads with performance creative. Three frameworks we use for every new campaign.
Server-side GTM — when it's worth it and when it isn't.
iOS 17, third-party cookie sunset, signal recovery — why half our clients don't need it, and why for the other half it pays off within a month.
Headless commerce: when it's worth it, and when it isn't.
Next.js + Sanity + Shopify Hydrogen — the modern e-commerce stack for smaller brands too. But does the SEO gain justify tripling the dev budget? A comparison across 3 real projects.
LinkedIn ABM campaigns in Latvian B2B — five experimental decisions.
Small audience size, high deal value, long sales cycle. Why the classic ABM playbook doesn't work here, and how we adapted it for real clients.
CRO audit — seven things we check in the first hour.
Before heatmaps, before A/B tests, before session recordings. Let's figure out whether your website is fundamentally broken.
llms.txt — does it actually work? An experiment across 12 client websites.
We measured whether LLM citation frequency changed over 3 months. We found an answer, but not the one we expected. Full dataset and methodology in the appendix.
A short note on agency work rhythm — why we don't work Friday afternoons.
A simple experiment we ran in August 2024 that changed how we plan sprints, client calls, and, paradoxically, execution quality.
Archive by year
2026 / Q2 · 12 articlesShall we turn this read into results?
If any of these articles made you feel like we speak the same language, let's start with a 30-minute call. No SDR gatekeeping.