Methodology and context of the analysis

[jahid12]

The analyses were conducted between October and December canada phone number list 2024, and the AIs are likely to evolve from then on. In simultaneous windows, we ran the exact same analyses against the AIs, loading both our own data from IKAUE (our agency) and anonymized client data. The tests included:
Loads of various Google Analytics 4 (GA4) data collections, exported from the interface, for which we request situational analysis and provide context on the strategies followed. Examples of campaigns, channels, behavior, and conversions.
Uploads of pre-downloaded and pre-organized product stocks, including sales data. We request stock, category, and product typology analyses.
Google Search Console data uploads to validate keywords and URLs, as well as their evolution.
Data copied from the browser to the results of various SEO tools: Sistrix, Ahrefs, etc.
Loads of the same data transformed into formats other than CSV: XML, JSON, HTML tables, and direct copies from Sheets and Excel.
These tests focus on the world of digital marketing. Based on these tests, I've drawn fairly general conclusions, as what has been seen is that there is an easily intuited pattern. When something doesn't work, it doesn't work in almost any scenario, and when it does work, it does so repeatedly in several of them. Based on this, and after discovering the benefits of the systems, but above all, the frustrating limitations they also have, I have created a classification of relevant aspects for analysis. This classification is entirely intentional and follows the initial tests to demonstrate what really matters, but as I said before, it is entirely subjective and very much tied to my personal judgment.
About me (in case you want to know)
Since the analysis is highly subjective, to provide more context for those who don't know me (if you already know who I am, feel free to skip it) , some details: I 'm Iñaki Huerta, a digital analyst and SEO. I entered the web world more than 20 years ago and have been in both fields for more than 17 years. I have been leading IKAUE for 10 years and working for a multitude of large clients. I come from the communications field, but I have been programming for as long as I can remember. Today, I program using PHP, JavaScript (& Node), R, and some Python (which is the one I'm least proficient in). I have a very high level of SQL and data processing, as well as its visualization in various tools. I have worked with Machine Learning and AI from the beginning and I am used to working with huge amounts of data. In short, a business-oriented professional with technical skills.
Categories of this analysis
The categories into which we divide this analysis are:
Data input and output: Here we discuss the system's capabilities to work regardless of the outcome. File formats, weights, ability to repeat requests, etc. The technical side.
Data manipulation and correction: How you first understand, organize, and prepare your data for analysis. That's the preliminary work every analyst has to do when the tool or data engineer doesn't give you all the work already organized.
Core analytics capabilities: How AI informs your analysis, understands your data, and helps you focus your research.
Data Visualization: The ability to provide graphical output that you can understand and use in your reports.
Gaining and explaining insights: How well it can give you valid information about the data that leads you to practical conclusions.
Below, we'll explore each of these categories in detail, evaluating how each AI performs and assigning a score from 1 to 5 for each subsection of the analysis.