Our Ranking Tech

Clausematics objectively ranks South African advocates who are officially listed with the constituent members of the General Council of the Bar of South Africa and/or who appear on the Advocates’ Roll kept by the Legal Practice Council.

In summary, we rely on:

  1. verified, real-world user input; and
  2. computation roles ran by our software script; then
  3. once processed, the data is analysed by the founding team using data processing and data visualisation tools.

The founding team comprises an attorney with distinctive, large-scale transaction and litigation experience and a software engineer with data science and AI expertise. The founding team is periodically assisted by esteemed members of the legal profession and software engineering industry on an advisory basis.

Clausematics iterates on the rank aggregation approach of a distributed, iterative algorithm. Our methodology builds on this foundation by applying the logic-steps set out below.

The Technical Details

User authenticity

  1. We receive a small fraction of “spam users”, suspicious accounts and “fake” user identities.
  2. We circumvent this via our verification procedure which will require a user to verify her/his identity by either (i) clicking a link sent to the provided email address; or (ii) a one-time-pin sent to the provided mobile number
  3. Our verification step randomly assigns these two measures, meaning in some instances an email link will be sent, and in other instances, a one-time-pin will be sent. The only data we deem legitimate is that which has been received from users who have been verified in this way.

Quantitative variables

  1. At this level, the user feedback is aggregated per advocate. The users assign a score to the advocate concerned according to the “vanilla” 5-star online rating system with 1 being the lowest score, and 5 being the highest score.
  2. Every user has an unlimited number of votes, provided that they will be capped at casting one vote per month for each advocate. In other words, a user may not vote for the same advocate twice in a 30-day period, but may vote for an unlimited number of different advocates once every 30-day period.

Qualitative variables

  1. Our hypothesis is that feedback is most effective when people are given the freedom to express themselves. This freedom must, however, be counterbalanced by its ability to provide meaningfol feedback.
  2. We have therefore devised a question which is both wide enough to capture the unabated thoughts of a user, while simoltaneously being specific enough to provide insight.
  3. We do not incorporate negative feedback in the ranking computation. Our focus is on the attributes which celebrate the outstanding abilities in people. Clausematics is not a complaints forum. We are a data-driven institution which aims to enhance high-calibre people’s ability to contribute to the world that which they are good at
  4. Members of the public wishing to lodge complaints are urged to use the well-established avenues of the Legal Practice Council, General Council of the Bar of South Africa (the “GCB”), the constituent members of the GCB being the Cape Bar, the Johannesburg Society of Advocates, the Pretoria Society of Advocates, the Free State Society of Advocates, the Society of Advocates KwaZolu-Natal, the Eastern Cape Society of Advocates, the Northern Cape Society of Advocates, the Polokwane Society of Advocates or the Mpumalanga Society of Advocates.
  5. We are still able to achieve the desired effect in our data despite discounting negative feedback in that mediocrity will simply have no measurable data points and therefore it’s non-recognition resolts in a lower ranking for advocates who do not receive positive feedback.

Weighted scoring model

  1. After conducting thousands of hours of analysis, we concluded that advocates oltimately serve to achieve the best legal outcome for a client.
  2. Clients, in turn, approach attorneys in the first instance. The outcomes obtained by an advocate, however, are often highly dependent on advocates’ ability to persuade members of the judiciary (and adjudicators such as arbitrators) of their point of view.
  3. There is also scope for great insight provided by anonymous and independent peer-review which necessitates the involvement of fellow advocates.
  4. The above logic informs our weighted scoring along the following categories of parties on a sliding scale:
Feedback party
Client (including a representative of an institution e.g. legal advisor, CEO etc.)
Judge/magistrate/adjudicator
Attorney
Members of the media
General public

all of the abovementioned (i) User authenticity step; (ii) Quantitative variables; (iii) Qualitative variables; and (iv) weighted scoring model informs the logic of our data modelling (“the Clausematics Rules”) which is then processed and analysed to produce our results.

Information processing

  1. We use a function-appropriate database provider as our batch processing framework which runs our received data and organises it using distributed iterative computation based on the Clausematics Rules.
  2. Thereafter, this batched and organised data is reviewed using:
  3. an intelligent text analytics software which is highly effective at Natural Language Processing; and
  4. a bespoke visualization tool which allows us to analyse and infer insights from the processed data.
  5. These tools work hand-in-hand and assist us in assessing the data received in a way it can be analysed efficiently and legibly,
  6. inclusively referred to as “the Information Processing”).

The Clausematics super-power

  1. Accordingly, the Clausematics Rules and Information Processing set out above allows us to (1) update the rankings in real time; and (2) Publish our monthly insights report which you can subscribe for HERE.

Future features

It is important that the core functionality highlighted above reaches a critical mass of data points to enable a solid foundation of accuracy. Once this has been achieved, we intend expanding Clausematics’ insights by adding the following:

  1. Practice area data: this will highlight which practice areas each advocate is strong at;
  2. Mass analysis of reported cases: our analysis will incorporate an analysis of courts and arbitration pleadings and decisions in which the advocate concerned was involved; and
  3. verification by big data auditing teams: for even more increased “hygiene” of the data, we will also include a quarterly comprehensive, independent audit of the data collected and processed by Clausematics by a reputable firm.

Please send any and all feedback on our methodology to thabu@clausematics.com