It was back in 2018 that we made the nature of our Matrix API known to the industry, although we'd used it internally for a number of years prior. With a pending SSL requirement imposed by Google back in 2018, we ran various assessments that assigned Sydney Broker Brokers with an SEO score, and more importantly, assessed how many of these brokers had a valid SSL Certificate installed on their website. A staggering 77% of brokers lacked the certificate, so we contacted industry and supported hundreds of brokers at no cost with the process.
UPDATE, 3RD AUGUST 2025: The level of interest in access to Matrix was unexpected, so we'll publish a couple of more articles detailing more common API requests, such as checking for duplicate posts, accessing post or page keywords, and a few more. What wasn't mentioned in this article (for brevity and relevance) is that the primary purpose of the get request mentioned shortly is to access or return a page that is not currently indexed. All requests for site crawling are actioned via this getendpoint, but it's not the primary response returned via the API when requesting page data. The general page response is detailed in an article title "Understanding the Financial Web with Matrix". The Matrix module plays a role in much of our site and page evaluations, but it also overlaps with other modules such as our Athena News (and RSS) Module.
■ ■ ■
What was previously known as the 'Industry API' has evolved into the Matrix API, and the service is heavily integrated into a number of our standard marketing modules, such as News (Athena), Xena, RSS, and SEO. This article introduces the simplest and most basic feature of the Matrix API - page analysis.
Matrix Introduction: This article has been sitting in draft for at least a couple of years. In the time since we created our Xena statistical engine, the Matrix module has developed in significant ways that prevent a full and complete writeup of API functionality. Rather than discussing the 40 'action' endpoints, we'll introduce only the most basic page-level functionality. Formal documentation will be manufactured when required.
The Matrix Name: When James Joseph Sylvester first coined the word Matrix in 1850 (described at the end of the article), he described it as “a rectangular array of terms, out of which different systems of determinants may be engendered, as from the womb of a common parent. At its core, a Matrix is a structured source, or a framework that holds data in an organised grid so that complex systems and insights can emerge naturally from simple parts. Your data is no different. Matrix is built on the same principle: we take raw, scattered, often messy information — pages, numbers, scraped content, site metadata, social signals — and organise it into a living, structured system. From this framework, you can generate meaningful statistics, reports, insights, and actions — all linked by clear, dependable logic. Just as Sylvester’s Matrix gave birth to entire branches of mathematics, our Matrix engine powers entire layers of digital intelligence: scraping, indexing, tracking, benchmarking — all from a unified, structured core. Matrix turns raw digital chaos into structured, actionable insight — just like Sylvester’s original vision.
What is the Matrix API
The purpose of the Matrix API was initially designed to crawl client (and other industry) websites in order to evaluate SEO effectiveness. It collects a large amount of page data, and follows links in a manner not unlike Google in order to establish a 'weighted' linking structure. From a broader industry perspective, we would (and still) crawl every known industry website in order to create digital snapshots that we use for granular statistical analysis. We always want our client websites to outperform others - and they usually do - but in understanding how competing websites were presenting themselves we could reshape client sites to perform more effectively than others. What we're left with after each crawl is a trove of data that gives us a distinct and seriously significant digital advantage over others that rely on guesswork or marketing rhetoric. While it might sound a little self-serving, I can't imaging a financial marketing company enjoying any real success with their digital efforts without the data we regularly assess for performance metrics.
Since the early version of the 'Industry API', the Matrix API has evolved into a very powerful and integrated solution, meaning that some of our Yabber modules now use the API for reasons other than SEO. The News (and RSS) module, in particular, use the API to shape an understanding of various pages used for social media or general display on websites.
The GET Action
The API supports searching of archived data by links, keywords, free text, image content (via AI), source, domain, and an extremely large number of other attributes. In demonstrating a basic get 'action' we return a large amount of single page data that tends to indicate how the broader archive itself may be queried, since any returned value may be used for filtering results of any type.
The standard RESTful Matrix endpoint is matrix/{action}/matrix.json, with a number of URL parameters applied as required (including the API Key).
The get action includes an id attribute with a 'Safe URL Base 64 Encoded' URL. The returned data queries the destination page, evaluates it, saves results locally, and returns most data.
What Data is Returned?
A standard get action query returns the following (if resolved):
title, Page Titleauthor, Page Authorhtml, Full Page HTMLexcerpy, Article or Page Excerptimages_general, Single Primary Page Imageimages_general, Images (general)text, Article Content (Text)links, Article Page Linksvideos, Identified Videos or Most Typesimages, Images in Article Contentlinks_page, All Page Links, Including those Outside Article Content. Includes nestedinternal,external, andimage_linksarrays.readability, Readability Assessments forflesch,flesch_kincaid,coleman_liau_index,gunning_fog_index,smog_index,automated_readability_index, andfry_readability. Each includes a bested array withscore,grade,age, andlabel. The 'readability' array also includes thesentences,number_words, andsyllablesfor the nested (inner) plain text article content.tags, Tags includes two arrays:glossaryandkeywords. The former is measured again your own (website) Yabber glossary, and the latter is an array measured against all inner plain text words but excluding common dictionary words. Each term array is presented asterm => count, with each array ordered bycount.meta, All pagemetatags of all types.screenshots, Screenshots of the supplied page. Returns in an array withoriginal,large, andthumbnail. This data is not typically served in the initial response to speed things up. Subsequent requests will include the data.json_ld, All Valid and Invalid JSON-LD SEO page markup. Each array includestype,valid, andjson. Invalid JSON is returned as a string.favicons, All favicons assigned to the page as links.social, Social links for most social platforms. Returned withplatformandurl.phones, All resolved Australian phone numbers of all types found on the page.emails, All resolved emails found on the page.credit_licences, Any Credit Licence found on the page.wp_version, If a WordPress website, the software version is returned.brokers, Returned on thebrokersactionendpoint. Returns an array of comparative data with a page score indicating performance when measured against all other industry websites. Thecompliancekey returns a list of AI-generated chnages that are required to meet legislated compliance requirements.
Other data is resolved but not returned. All data is quickly evaluated, measured against other pages, and appropriately indexed into the archive for other queries.
Favicon API: Favicons also operate autonomously under the Favicon API. Given that clients generally use our [link] shortcode to render external (and internal) links on their website (placing a destination page icon alongside outgoing links), we also measure the count of each icon returned alongside the link. Since we return external links in Matrix, the data isn't super useful but it does give us another means to rank external websites. It was more useful when we had a free plugin to manufacture links. However, outside of Matrix, the API can be used to return just the Favicon data.
We'll query Westpac.com.au. The result below (heavily truncated and snipped for readability) shows a partial response. Page HTML and text content was excluded for length.
Only data that was found or resolved is returned, but you'll get a feel for how the JSON is structured when you query a few pages and unwrap the response. In the above example there was only limited linking data returned and other information simply wasn't found. When found, the data is returned in an array.
We return a large amount of readability data, but it's the flesch_kincaid score that we use on website (for SEO analysis).
The credit_licences array is greedy and all matching data is returned on the initial query. Subsequent queries will return only those licence numbers that validated against the ASIC API.
The links and links_page response differs in that the latter evaluate the entire page, so the header, footer, sidebar, and everything else, while the links array returns only that data found in the general 'article' content, and it's the latter we use for the majority of balanced internal linking scores.
Client Website Links: For client websites, we resolve internal linking to a page ID in order to create a page Matrix. Client websites are indetified by a site_id in the query along with the apikey. Slightly more information is returned for client websites for the purpose of feeding the Xena statistical engine - particularly as it relates to linking - although the response is normalised for the purpose of comparing against other industry websites.
In crawling a single website we're able to extract a meaningful and very structured linking structure that builds a Site Matrix Architecture (mirroring Google's weighted PageRank algorithm), and we're able to grade the website or identify areas of weakness, but the power of Matrix comes in our capacity to compare one website against others for broader statistical purposes. The more complex data isn't returned in the standard API response because it's lengthy and largely irrelevant, but you may return that data with matrix=yes in the URL call.
Xena Statistical Engine
We've built a proprietary statistical engine called Xena that provides holistic and aggregated marketing statistics to users. Resolved to the user level, the system replaces an array of (legacy) standalone siloed tools to provide web, link, download, mouse movement, action, heatmaps, and other functionality. Very heavily integrated with our BeNet AI engine, the system isn't one that has really found its place in the mortgage market (which is very slow to adapt to new technology or ideas), but it arms us with the vital statistical data to craft website pathways, funnels, and ultimately deliver better results.
We regularly wash statistical data against Xena to build highly accurate reports for the purpose of identifying precise human behaviour, and our resolved understanding is the foundation upon which our (recently updated) landing page module relies (and why our landing pages and methods tend to change so often).
Xena is more effective than any competing statistical module, and it's a module that we're particularly proud of given how derived insights have empowered us to deliver even better results.
Another self-serving closed statement: we can categorically state that if you're not using Xena, you're not getting the most from your marketing efforts. There are commercial solutions that come close, but they'll cost you hundreds per month.
Yabber Beta: If you're using the Beta versions of Yabber, you'll notice that Xena has disappeared completely. The icon in the header now links to global statistics, and the page will return the full suite of Xena data very soon. Legacy users will find statistics via the 'Website' module.
News 'Athena' Module
A News Aggregation module - commonly used for website news articles or social media content - called Athena (the Greek goddess of wisdom, strategy, and knowledge) uses Matrix natively to assess every single page (from an RSS or other feed) that is indexed by Yabber. Weighting the news terms against your own glossary and dictionary terms empowers you to quickly identify news trends but also identify articles that are most relevant to your audience. Having Matrix filter, categorise, and prioritise news articles gives you immediate access to all the page objects and data necessary to identify relevance and shape an agile social news strategy.
Athena provides insights that feeds us with 'trending' information that generally translates to consumer sentiment (evaluated via the 'Sentiment API' which overlaps the Matrix and Xena ecosystem). In assessing language, news, the direction of aggregate opinion, and even frequency of terms, we're able to apply the agile strategies on Facebook (advertising platform) or elsewhere that is consistent with emerging news cycles and a collective consumer conscious. Not discussed in this article, the terms API action with a frequency=daily parameter is super simple, but also staggeringly accurate in articulating emerging consumer trends.
Insights API: To confuse the matter even further, the Insights API accesses industry advertising and conducts and algorithmic and full AI analysis. The data from this API is measured against others to determine how advertising matches up with Sentiments and known trends. It also conducts a full compliance audit.
Outside of the scope of this article, we've only just released an enhanced Athena News module in Yabber. Once a hidden module, we'd like to see it used for a range of purposes that we'll introduce in our FAQ module.
Broker Insights
Matrix has evolved enormously since we introduced the early versions over 15-years ago, but we still apply a very clear focus on those formative objectives - measure the 'Financial Web'. That is, to collect and compare all known broker websites against each other in a way that allows us to rank each site against each other for compliance, performance, and activity.
We 'Open Sourced' the Matrix broker data last year and provide deidentified data dumps for those that want it. However, that position has changed a little and we now require registration and approval. Those with an API Key will always have full and complete access.
As we've done for the last 15-years, we continue to publish a yearly paper that exposes digital trends on all known broker websites.
To access only broker data (and exclude the millions of non-broker websites), use the action of brokers with applicable URL parameters to filer the paginated results. Unlike other actions, accessing the broker/matrix.json endpoint directly (with no URL parameters with the exception of apikey) returns all available endpoints, parameters, and options - essentially a Swagger file.
Why Our Broker Website Performs Better
The result of any website is precisely zero if not promoted or invisible in search, and this overarching truth has to be understood. Having the greatest digital or social presence returns nothing if you don't have an audience. What we can categorically say is that our broker website framework outperforms others on the market, and we reach that determination based on assessments that others ignore. We include a range of features and tools, gamify the user experience, provide more education, natively integrate assets such as video, and manufacture intelligent pathways and internal funnels via a choice architecture shaped by a statistical and real-world understanding void of fluffy guesswork. The entire broker website is crafted based on known performance metrics.
We take a vastly different approach with our Facebook campaigns, so the success of our brokers has a lot to do with integrated funnel-centric technology that is unique to our experience, but the campaigns themselves are often determined by the data we've resolved from your users or industry sentiment - an extremely agile method that has consistently enabled us to return better results.
One Million Challenge: For over 6-months we've run a milion-dollar challenge. We will multiply the results of existing Broker Grow, Transformers, Bizleads, and a bunch of other really low-performing generic solutions or give you $1m. In reality we've generally improved results by 500% - 800% without effort - sometimes better... and usually with minimal effort. There are a number of reasons for this, but our insights certainly play a part.
Conclusion
In an industry where digital presence is often reduced to surface-level aesthetics or the convenience of an off-the-shelf template, we remain unflinching in our belief that data is the true currency of competitive advantage. Our Matrix ecosystem — supported by Xena’s powerful statistical intelligence and Athena’s curated news insights — is not simply a passive tool but an active engine that fuels every informed decision we and our clients make.
The reality is that the financial web — and brokers as its custodians — are entrusted with more than just loans and lending products; they shoulder the responsibility of public trust. It is this trust that must be earned and sustained by rigorously applying accurate data, real-time insights, and actionable intelligence. A headline is never just a headline. A link is never just a link. A page is never just a page. Behind each digital element lies an opportunity to understand, adapt, and outperform — and Matrix ensures those opportunities are never wasted.
While many speak about 'data-driven marketing' as a slogan, we built a framework where this ethos is embedded into every pixel, click, and funnel. We are not marketers who dabble in statistics - we are statisticians and analysts who engineer marketing systems. This is our edge, and it is the edge we hand to our clients daily.
We will continue to invest in the relentless pursuit of better data, clearer insights, and the systems that transform that intelligence into meaningful outcomes. The Matrix API, Athena’s watchful curation, and Xena’s surgical precision are not just tools in a toolkit — they are guardians of your digital advantage.
The level of digital and analytical clarity we've just described will outpace those who still operate on assumption, stale spreadsheets, or empty promises. This is not marketing hype — it is measurable, repeatable fact that undeepins the Financial Web. If there is one truth to take from this: those who measure more, know more. And those who know more, win more.
History of The Matrix Term: James Joseph Sylvester (1814–1897) formally introduced the word 'matrix' into mathematics in 1850. He used it to describe an array that gives rise to determinants: Sylvester’s original wording (1850): “I propose to term...a rectangular array of terms, each of which is a function of the same set of variables, a Matrix...”. (Sylvester, “Note on the Algebra of Matrices”, Philosophical Magazine, 1850). Arthur Cayley later expanded this work in his 1858 paper, A Memoir on the Theory of Matrices, which laid the foundation for modern matrix algebra. Etymology and Meaning: The first use of the term Origin: Latin matrix “womb, source, mould” (from mater “mother”). The first known academic use of the word was in c1374 when Geoffrey Chaucer, in his Treatise on the Astrolabe, uses “matrix” in the sense of a “womb”. Early 15th century medical manuscripts and surgical treatises use matrix for the womb, borrowed directly from Latin medical Latin texts. The earliest citation for the meaning “mould in which something is cast or shaped" (essentially laying down the foundation for the modern interpretation) was in 1555 when Phaer Virgil quoted that "Like as a bee in matrix wrought". By 1656, Thomas Blount’s Glossographia defines matrix explicitly as “the mold in which letters are cast". Our assignment of the Matrix name comes from Sylvester's "an array that gives rise to determinants" definition. Before Sylvester, mathematicians mainly talked about determinants directly — the value was the focus. Sylvester’s insight was to separate the idea of the array itself (the structure) from the number it generates. He gave the structure its own identity — the matrix — and this opened the door for modern linear algebra, where we add, multiply, invert matrices ... thinking of them as objects in their own right, and using them to represent transformations, systems, and more.
