I’ve grown accustomed to a three pronged approach from media in terms of coverage for New Zealand tech companies;

  • Sam Morgan and TradeMe get warp level coverage across tech, business and general channels
  • Rod Drury and Xero get pretty decent coverage in business and tech channels
  • Tim Norton and PlanHQ (Tim – learning well from his mentor Rod) – get to pick up the tidbits with pretty good tech channel exposure

So, I was stoked to see this morning that another player, and a Christchurch one at that, got some great exposure. VortexDNA is a cool little (or not sol little) company that seeks to monetize its technology for measuring and predicting human intention. It has created a pattern of human intention that it beleives is akin to DNA.

Well done Raf and Branton and I’ll be interested to watch the development of your product(s) and business.

Ben Kepes

Ben Kepes is a technology evangelist, an investor, a commentator and a business adviser. Ben covers the convergence of technology, mobile, ubiquity and agility, all enabled by the Cloud. His areas of interest extend to enterprise software, software integration, financial/accounting software, platforms and infrastructure as well as articulating technology simply for everyday users.

  • Falafulu Fisi |

    Stuff article says…
    It allows predictions to be made about people and who they are, which could have huge ramifications for the insurance and advertising industries, says VortexDNA director Branton Kenton-Dau.

    I think that VortexDNA needs to benchmark their application against other popular & commonly used predictive modeling techniques that are currently in use today. The reason for them to do so is to give credibility to the claim of the product. I mean what is new in VortexDNA that is not already covered in current state of the art techniques such as kernel-machines and random forests, which there are many variants of them available today (each one is different only in its classification error rate, some are lower than others). If VortexDNA can show that it has a lower classification error compared to other techniques, then it is good for them in terms of marketing. The other reason for them to benchmark is to perhaps convince more VCs that theirs is superior to the current models.

  • Great stuff. I hope when they become the billion dollar company they aspire to be that they are still based in Christchurch

  • Thanks guys for your support. We certainly plan to keep Christchurch as our core base even as the business expands overseas.

    Thanks Falafulu for your comments. We are driven now by the needs of our customers and the marketplace. VCs can have you jumping through hoops for many years trying to prove something is perfect. It never is. Sometimes you just have to get out there and engage.

  • Falafulu Fisi |

    Raf, the reason of my comment is because predictive modeling is my specialist area and from the description of your product it is claimed to be many things, such as complex adaptive systems, well what does that mean for the average customer that you’re trying to sign up? It means nothing. Besides, any potential customer out there whom probably hire an analyst to evaluate your product before they buy a license will find out soon that your claim is not what the term complex adaptive systems is. It involves emergent dynamics & non-linearity, which I am yet to convince that your system is able to do. The reason is I’ve yet to see a commercial system that is labeled as a complex adaptive systems. See, if you get found out by an independent consultant (of a potential customer of yours) that the product is being hyped, then you will lose some credibilities when you try to market it to your next customer.

    I had been involved with the expert group that drafts the Java Data Mining API version 2 (JDM-2.0) for the official Java technology, and some of the Vendors involved in drafting this specs, are the top in predictive analytics which they dominate, SAS, SPSS, KXEN, Oracle, IBM. I also know that some of these will move big time into the area of predictive risk modeling, in which is an area that VortexDNA wants to get into , ie, Insurance Industry. Take for example, Fair Isaac Corporation (which they also involve in drafting JDM 2.0), predictive analytics software is widely adopted in the insurance industry. Fair Isaac’s R&D team kept improving/inventing their algorithms on a regular basis, and I know this since their 2 members in the JDM 2.0 expert group are both academics who help teaching positions at University. See, if VortexDNA & Fair Isaac do submit RFPs from a big Insurance company, and both your technologies are required to be demoed , I believe that Fair Isaac will come on top, this is not to mean that VortexDNA has no chance, but based on my experience in knowing the academic backgrounds of 2 Fair Isaac members in our JDM 2.0 expert group, ie, what they understand about predictive analytics, and what’s their vision for the future. I even learnt new things/topic in our expert group online discussion for feature proposals for the JDM 2.0 . Even if you secure a customer , it won’t guarantee that this customer won’t defect to your competitor if they can show that their system is more robust than yours.

    In high risk business such as insurance, analytic technology selection is mainly tilted towards one that has a lower-error classification rate, because a difference of an error rate of say, 0.001, between 2 vendor’s analytics (perhaps vendor A has error rate of 8.7 and vendor B has 8.699), they (insurance vendor) will likely to go with the analytic vendor with error rate of 8.699, the reason is that a small difference like that could be the difference between losing a few millions and gaining nothing. This is why I think that you should benchmark your tool to show potential customers of why they should buy licenses of your tool.

    If you say, that you’re driven by the needs of our customers, then I think it is being naive, in that you think that your existing customers will never get a visit/phone-call/email from sales reps of your competitor, say Fair Isaac. You customers will defect , the minute they see a demo of a product that is superior to yours. Even if it is not an existing customer but a potential customer that you’ve been talking to for some weeks/months, once this potential customer has seen a product of a competitor of yours that perhaps superior, then they will like to terminate all contacts with VortexDNA, since they’ve found what they thought is best solution for them.

    There had been 2 recent cases of NZ companies , where one that just closed shop , Compudigm International (after 10 years in business) and and one that was sold off, ie, Esphion (after 5 years in business) that there was hardly any profit at all to the original investors. These 2 companies are in the same domain as yours , ie, predictive analytics. It is widely known that Compudigm found it hard to expand their market from Casino’s to anything else, and because I knew the algorithm they based their technology on, which is SOM (self-organise-map) neural network, I am not sure if they ever improve the classification error of their SOM based engine or not, but one can only infer why they couldn’t break into other market apart from Casinos. On the other hand, Esphion, had demoed their network security to lots of US potential customers including the US Defense Department, and again they found it hard to gain markets over there. See, any demo of such product to the US Defense Department, they get experts from Los Alamos to evaluate the technology on their behalf. Los Alamos scientists are expert in almost everything predictive modeling, since that is what they do in their daily job at Los Alamos, and they can tell superior technology once they see one. Esphion uses Neural Network as its core engine for intrusion detection and there are about 60 variants of different neural network available today. So, I wouldn’t be surprised if another vendor outperformed Esphion for the Defense Department contract.

    I strongly advise that you must hook up with R&Ds at your local univsersity.

  • Thank you for you in depth advice.

  • I didn’t know I PlanHQ were getting that much action : )

    I am just learning, and look forward to seeing heaps more coverage of more companies in the media, its one of the areas we just all need to get better at.

    Good to meet you in person super quick at the event last week, didn’t have a chance to catchup.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.