What challenge is PROS helping businesses to overcome?
One of the big challenges they have right now is how to drive sales growth and profitability improvement. They want to mine operational data and external data, in order to know which accounts to go after, and drive more confidence in their selling.
In a lot of companies the sales organisation has lost a lot of confidence in negotiating. Procurement groups have professional teams, and they’ll tell you that all products are the same and that your price is much higher than the competition. Sales need real-time tools to better guide their price points. We start with basic master data and transactional data from the likes of CRM and ERP, then we put that on our platform to connect other data and make it real-time.
What sort of other data feeds do you typically pull in?
It could be returns data, or warranty claims data. Lots of data is disconnected from the customer, but if you can reconnect it you can drive a prescriptive action. For example if you are seeing abnormal product returns the product guys can look at that or account managers can call the account to finds where there might be a problem.
Can you offer any other examples?
The demand for natural gas is affected by weather and temperature. You can use that external data as leading indicators of product demand patterns. Another example is car service points: you can look at new car registrations, the age of cars and so on to predict which parts are most likely to need replacing and which materials you are most likely to need. It’s about predicting growth opportunities and aligning sales and inventory to where it’s needed most.
It sounds a bit like using any other analytics. So what does PROS do that I wouldn’t get from more ‘vanilla’ analytics?
Standalone analytics do not drive change. You have to go from just having analytics to actually driving an action. Ours is real-time, so the analytics are tied to execution, so change happens automatically. An analogy I use is that cars used to beep when you reversed towards an object, but the latest systems will actually stop the car when it detects objects are too close.
The other thing is that we build around a specific challenge, for example price management and price optimisation. When something changes, like there is a currency fluctuation, we automatically re-price all the products that are affected.
And how does it differ from complex event processing [CEP]?
We built our own real-time engine and our own forecasting algorithms. We built algorithms for linear and non-linear outcomes. The algorithms are part of our secret sauce. We also have outlier detection – for example there may be outliers that are not really representative. For example if you have a particularly high number of cancellations one day due to a freak storm, you may not want to use that to predict the future. A one-time event might need censoring out.
What do you make of the whole big data hype?
Unstructured analytics don’t drive any value. If you can’t drive a specific action then it’s just interesting. You need to connect back to your operational data. We think we can add two more ‘v’s to the big data question: value and viability.
Do you have a single platform or different apps for different use cases?
One of our biggest strengths is that we have one platform for 35 industries. You then have configuration templates for specific industries to bring in best practices in that space.
Is there much customisation work required?
Implementations average from six to nine months. We use implementation partners to help but some customers have done it on their own with some training from us.
You recently announced an OEM agreement with SAP to embed the SAP HANA platform with your big data applications. What does that give customers?
HANA gives you not just real-time but also the R language and built-in analytics that we can tap into at the data access layer. 50-60% of our customers are also SAP customers so it makes sense. We’ve been investing in that for six months, and it runs today, and we expect to reach general availability in the second half once we have certification.
You’ve talked about brining other data feeds into your platform as events and triggers. What sources can you ingest?
We have built-in integration of the likes of SAP, Oracle, Microsoft and salesforce.com. We’ve also built some connectors that are Excel-based.
Is the application cloud-based or on-premise?
It’s a web-based app so it can be deployed in a private cloud or on-premise. We also have a native Force.com application for salesforce.com customers. The app can sit on Windows or Unix servers. If companies are dealing with multi-Terabytes or Petabytes then for performance we rely on Violin Memory [flash storage] arrays.
You call yourselves a big data company. Is there any risk that big data might not take off as hoped?
I don’t think so. The reality is that customers who are not using the data available to guide their business will be at a disadvantage. The airline industry adopted the idea of big data from early on. They forecast based on holidays, the weather, external events and so on. They have had to do that, because 70% of their costs is fuel and they operate on very thin margins. Otherwise they can’t survive. The fact is that we’ve been doing big data since before the term was coined.