Predicting consumer behaviour with tech | Fin24

Predicting consumer behaviour with tech

Sep 27 2018 13:02
Glenneis Kriel

It can take up to 18 months to write artificial intelligence (AI) for specific businesses. 

Xineoh is developing standardised programmes that can be adapted for a business within four weeks, with results that compare well with the best international tailor-made solutions. 

Vian Chinner has been working on the solutions since 2009, first at his company TrafficSquared and now at Xineoh. He explains how it all works.

Tell us about your background.

I grew up in Bloemfontein and studied financial mathematics at the University of the Free State. 

I started my first business at the age of nine, selling toffees and suckers from a lunchbox. 

This later expanded to velcro wallets and lip ice. 

My profit was around R1 600 a month in today’s terms. Then the school put a stop to it. 

I didn’t realise it at the time, but that was my first exposure to how much control government and the authorities have over entrepreneurs, especially here in South Africa.

What did you do after school before you started Xineoh?

I interrupted my studies in my third year and joined a hedging company in Stellenbosch, but did finish my degree at Kovsies later. 

After a year, I started doing commodity arbitrage on my own – until liquidity in resource markets dried up in 2009 because of the financial crisis.

Together with a friend and family members, I invested the capital from my arbitrage work and started TrafficSquared, a company that specialised in the development of AI for the automatic buying and reselling of advertising space on the internet.

TrafficSquared was bought out in 2013 by a company in Los Angeles called Ylopo, which was funded by Move – the operating company for, a large property search portal similar to Property24. 

Ylopo developed AI that ran Realtor’s recommendations engine. It really was a great project and we achieved quite a lot of success with it.

What exactly does Xineoh do?

We enable companies to predict consumer patterns within a few weeks, such as which product will be bought by whom at what price and at which level of demand. 

None of our competitors can implement such a system at the same speed as we can.

How did you manage to develop such a product?

I think we were lucky because at first I could not believe how accurate our predictions were. 

It also helped that we use the technology ourselves and did not have the luxury of developing technology for the sake of developing technology.

Why and when did you start Xineoh?

I worked for Ylopo for about 18 months before I started Xineoh in 2014. It was not a whole new start. 

The entire team of programme engineers who worked for Ylopo in South Africa, myself included, broke away.

Our long-term vision for the industry differed too much from Ylopo’s.

My vision is based on the realisation that AI can never be broadly applied if every solution has to be written afresh for the unique features of specific companies. 

Standardised products are needed to make the technology more accessible to all.

Why the name Xineoh?

Xineoh comes from the word phoenix spelt backwards because Xineoh, like the phoenix, apparently can’t die. 

After every problem or crisis, we emerge on the other side alive and stronger.

Where did you get funds to start Xineoh?

Xineoh started as TrafficSquared in 2009. I got funds from various sources, including self-funding, funding from friends and family and angel investors.

Last year, I also acquired venture capital worth $2m from an investment bank in Vancouver, King & Bay, to expand the platform.

How has your market developed over the years?

To prove the product, we initially worked on commission for Amazon and eBay.

Approaching Amazon was as simple as sending an email, but the American eBay took no notice of me for years, until I finally contacted the right person. In 2015 they endorsed us for an American eBay contract.

Amazon and eBay could probably have generated good revenue for us for the next decade, but they were too big, and repressed any further innovation. Last year I decided to target companies directly.

How did you market the product to companies?

I attended conferences in the US and that’s how I made friends. It wasn’t easy because I’m an introvert and they don’t really understand the Afrikaans accent. 

I realised later there are some super extroverts who sometimes refer to themselves as “connectors”. 

It helps to become part of their networks, because they can introduce you and get you in where you should be – usually at a referral commission of around 5% to 15%.

What have been your biggest challenges so far?

Failures and rejection by customers, investors, suppliers and prospective employees can actually get you down a bit in the beginning. 

And it’s also a lonely road – you’re not really part of a team. You are the team. To get the right partner can help, but that’s easier said than done.

Also, you must guard against burn-out. At 35 you can’t pull all-nighters like you did in your 20s – you have to start looking after yourself.

I found finding the right staff is the biggest problem. 

I don’t have a perfect plan yet, but I have learned that it is better to keep a position vacant than appointing someone who isn’t entirely suitable for it.

How has your business grown since you started it?

Xineoh was just me. We are now 12 people. We have also been much bigger, but every person in the company must be relevant.

How does the South African market compare with the international market for these services?

For big South African companies it’s much the same as for large American companies. Everyone realises the importance of implementing AI. 

Accenture estimates that the average company’s revenue could grow by around 38% by 2022 thanks to this technology.

However, the supply of engineers who can put AI systems in place at pioneering level is extremely limited. 

Element AI estimates that South Africa currently has fewer than 100 such engineers. 

America has 9 000, but 4 000 of them work at one of four companies, Google, Facebook, IBM or Microsoft. 

There is a huge market for such services in South Africa. 

Each company will eventually have to acquire such a platform or build one itself.

The market is currently at the level we were on with computers in the 70s.

What other challenges are being experienced in South Africa?

Taxation is huge compared with many other countries. It destroys capital formation.

Economic prosperity is often blamed as the enemy, but it’s conveniently forgotten that almost every successful business needed venture capital at some point. 

Venture capital comes from economic surpluses, or put differently: wealthy people. Without them there is no venture capital, no new businesses, no pioneering work and no new jobs.

Currency control makes it extremely difficult to do international business and has already jeopardised more than one favourable deal for Xineoh. I don’t know why exchange control is still needed these days.

Another challenge is labour legislation – it’s definitely written by people who have never tried to run a small business. Every fringe benefit increases the price of labour.

What surprised you?

To tell you the truth, every success is a big surprise. In business, 30 things go wrong for every one thing that goes right. You survive all the things that go wrong, because as a small company you can quickly change direction.

What is the best business advice you have ever received?

In the US, the Republicans are fond of saying “a rising tide lifts all boats”. The best thing you can do for yourself as an entrepreneur is to choose the right industry that grows quickly, and where there is no dominant competitor yet. That wind at your back will make up for many of your other mistakes.

I am also a stoic – I believe you control your own thoughts and deeds. Channel all your focus on what you want to do, and accept the things you can’t control. 

This gives me the drive to get through the difficult times.

What are your goals for the next five years?

That Xineoh is either acquired by a larger company or is listed, probably in Canada. 

We will soon start another financing round and open full engineering and sales offices in Canada. 

We are also opening an office in Johannesburg in the next few months and will eventually move the core of the business from Bloemfontein to Johannesburg. 

This article originally appeared in the 27 September edition of finweek. Buy and download the magazine here or subscribe to our newsletter here.

technology  |  consumers

4 June issue
Subscribe to finweek