[{"data":1,"prerenderedAt":123},["ShallowReactive",2],{"blog-data-analysis-for-growth":3,"blog-related-\u002Fblog\u002Fdata-analysis-for-growth":92},{"id":4,"title":5,"author":6,"authorRole":7,"body":8,"cover":75,"coverAlt":76,"date":77,"description":78,"draft":79,"extension":80,"meta":81,"navigation":82,"path":83,"readingTime":84,"seo":85,"stem":86,"tags":87,"updated":90,"__hash__":91},"blog\u002Fblog\u002Fdata-analysis-for-growth.md","Data analysis for growth: from spreadsheets to decisions","William Ali","Founder & AI Specialist",{"type":9,"value":10,"toc":68},"minimark",[11,20,25,54,58,61,65],[12,13,14,15,19],"p",{},"Before models, before agents, there is a question worth answering: ",[16,17,18],"em",{},"what does our data already tell us that we're not acting on?"," A focused exploratory analysis often delivers more value in four weeks than a six-month AI programme, and it produces the clean, understood dataset that any later model will need.",[21,22,24],"h2",{"id":23},"what-a-good-analysis-engagement-looks-like","What a good analysis engagement looks like",[26,27,28,36,42,48],"ul",{},[29,30,31,35],"li",{},[32,33,34],"strong",{},"One decision in scope."," Budget allocation, churn drivers, pricing, capacity: pick one.",[29,37,38,41],{},[32,39,40],{},"All the sources in one place."," Exports are fine at this stage; a warehouse can come later.",[29,43,44,47],{},[32,45,46],{},"Honest write-ups."," What the data supports, what it doesn't, and what would be needed to know more.",[29,49,50,53],{},[32,51,52],{},"A dashboard the team keeps."," Built around the decision, not around every available metric.",[21,55,57],{"id":56},"a-short-example","A short example",[12,59,60],{},"A D2C brand had three platforms and three different revenue numbers. Four weeks of consolidation and cohort analysis showed which channels actually converted; reallocating spend lifted conversion by roughly 18% the following quarter. No machine learning was involved, just clean data and clear questions.",[21,62,64],{"id":63},"when-to-reach-for-ai","When to reach for AI",[12,66,67],{},"When the analysis reveals a repeated prediction or classification that people are doing by hand, you have found your first model. When it reveals a repeated multi-step process, you have found your first automation. Analysis first makes both faster and safer.",{"title":69,"searchDepth":70,"depth":70,"links":71},"",2,[72,73,74],{"id":23,"depth":70,"text":24},{"id":56,"depth":70,"text":57},{"id":63,"depth":70,"text":64},"\u002Fcovers\u002Fblog\u002Fdata-analysis-for-growth.svg","Line-art illustration of scattered data points on a chart resolving into a rising trend line with an arrow.","2025-02-05","Why most companies should start their AI journey with rigorous analysis of the data they already have, and how to do it in weeks.",false,"md",{},true,"\u002Fblog\u002Fdata-analysis-for-growth",5,{"title":5,"description":78},"blog\u002Fdata-analysis-for-growth",[88,89],"Analytics","Strategy",null,"xdDxzxzcjLhOACqip-Usi_LQ_cXYEyBLS_XomIA4t1o",[93,102,113],{"path":94,"title":95,"description":96,"date":97,"tags":98,"cover":99,"coverAlt":100,"readingTime":101},"\u002Fblog\u002Fmeasuring-ai-roi-honestly","Measuring AI ROI honestly: a method that survives the finance review","How to measure the return on an AI project without fooling yourself: baselines, full costs, holdout attribution, leading indicators and inflated-ROI red flags.","2026-08-11",[89,88],"\u002Fcovers\u002Fblog\u002Fmeasuring-ai-roi-honestly.svg","Line-art illustration of a ruler measuring a stack of coins beside a rising bar.",7,{"path":103,"title":104,"description":105,"date":106,"tags":107,"cover":111,"coverAlt":112,"readingTime":101},"\u002Fblog\u002Fmultimodal-ai-business-workflows","Multimodal AI in real business workflows: where it earns its place","Vision-language and audio models now read forms, photos, screenshots and calls directly. Where multimodal AI beats OCR-then-LLM, with examples, limits and cost.","2026-06-30",[108,109,110],"Computer Vision","LLMs","Automation","\u002Fcovers\u002Fblog\u002Fmultimodal-ai-business-workflows.svg","Line-art illustration of an eye, an ear and a speech bubble converging into a single processing node.",{"path":114,"title":115,"description":116,"date":117,"tags":118,"cover":121,"coverAlt":122,"readingTime":101},"\u002Fblog\u002Fmlops-basics-first-model","MLOps basics for teams shipping their first model","The minimum MLOps a team needs to put a first model into production safely: versioning, reproducible training, a registry, serving, monitoring and rollback.","2026-05-19",[119,120],"MLOps","Product","\u002Fcovers\u002Fblog\u002Fmlops-basics-first-model.svg","Line-art illustration of a looping pipeline with gauges, from data to model to monitoring and back.",1792973686218]