Lab – 01

kWINtel.

kWINtel is the keyword research and prioritisation tool built for my own client work, because most keyword tools will tell you volume and difficulty but not which of those keywords are actually worth your time this quarter. They also lacked the ability to define language and semantics which is my preferred approach to what I believe is one of the most critical things in marketing : understanding your customers.

It's the research and prioritisation framework that identifies genuine opportunity, understanding a business's existing and potential customers, the language they use, and how that language is interpreted by search engines and AI systems.

It is an internal tool rather than a product – it runs locally, it is not sold or licensed, and it exists because it is what I reach for on live client work. It gets extended whenever an engagement needs something it cannot yet do. It also helps that this is something that would used to take a solid week to do on your laptop a decade ago now is largely automated from a data perspective it still requires me to make decisions on setup, rules, configuration, language adjustments etc so I am able to get to the gaps, learnings and recommendations faster!

It also factors in how those same queries are interpreted by AI systems – not just where a keyword ranks, but whether it's the kind of query that surfaces in an AI-generated answer at all. Very useful in GEO too for identifying topics and questions you should be potentially optimising for and tracking.

See it in action

What it does

Eight jobs, in the order they happen. How each one works is the part I would rather show than write down. A visual the walkthrough below is the tour.

  1. Keyword ingestion
  2. Rule-based classification
  3. Persona & funnel mapping
  4. Gap analysis
  5. Content recommendations
  6. Query enrichment
  7. Visualisation & reporting
  8. Exports & brand isolation

Inside the tool

A walk through the pipeline, from a raw keyword export to a prioritised view of where the opportunity actually sits.

Set up Every engagement starts as its own project, so keyword universes, rules and taxonomies never leak between clients. Also handles multiple websites under a single brand, ccTLD's etc.
Seed builder Seed terms are normally one of the starting points when doing doing Keyword Research in SEO. Here we are also making use of match type filters to broaden or constrain our seed terms as required.
Import Semrush exports drop straight in as .xlsx or .csv, several files at a time. Designed to work both manually and via API's.
PAA explorer Fetches People Also Ask questions for up to fifty keywords at a time, with a sandbox mode for testing endpoints and what type of data you'd normally expect to receive, without spending API credit. Very handy when building out automations and workflows!
Data cleaner Scans the active keyword set for duplicates and missing classifications. Nothing is changed automatically as this was built for cleaning datasets without needing to open databases. Every finding is reviewed and actioned individually. Some of these can also surface the need for additional rules or classifications you might want to include.
Classification rules Sub-categories are driven by pipe-separated trigger terms with scope, priority, brand overrides and exclusions. A rule can be tested against a single keyword before it is committed, or the whole set bulk-imported by CSV.
Classification summary 417,681 keywords resolved into 54 sub-categories, each one traceable back to the rule that placed it. Far more keywords than your average UK business would need to track, this tool was rather built out of frustration with a certain online spreadsheet app constantly freezing with larger datasets.
Clean topics 216 topics mapped to sub-category, bucket and funnel stage, filterable from awareness through to loyalty. I also decided to build in a way for my AI of choice to help with this step as it is just data crunching and comparing.
Taxonomy & personas Buckets and personas layered over the classified set, so the same keyword universe can be sliced by who is searching rather than only by what they typed. Additionally this also gives us a visual way to start better understanding the end customer, their persona, the language they use and which stage of the funnel they really are.
Personas & bucket mapping Personas are defined per project, and every sub-category is mapped to a bucket and a funnel stage. Not designed to be too rigid however, as almost every project will have a range of differences and your customers are no different!
Project config Domain and API settings, held per project.
Data explorer Free-text search across keywords, GEO topics and clean topics, filtered by sub-category, funnel stage, intent and volume, with sub-categories excludable on the fly.
Insights A variety of insights that we can filter by sub categories and drill into the data views we need across whichever products or services we were looking into.
Volume by sub-category The biggest clusters ranked by total search volume. A great starting point when looking at product or category performance and could also be used for making decisions on which areas of content they should look at expanding.
Seed map Every approved seed with its match type and a note on why it earned a place.
Reporting The strategic search landscape in one view, assuming we did a good job with the other steps! Demand split across the buyer journey, and the biggest keyword clusters ranked by volume. It would also not have been me if it did not include some 'options' by way of Charts - and there are LOTS out there!
Gaps & personas Coverage status and largest uncaptured volume by sub-category, then the same demand re-cut by persona and by where each persona sits in the buyer journey.
Manual mode API calls can be switched off entirely, falling back to Semrush exports and manual uploads. It is also always wise to keep a log when API's are involved!
Admin Sidebar sections can be shown or hidden globally, across every brand at once.
01 / 19
Ask me about kWINtel →