Why enterprise data analytics platforms haven’t displaced bespoke solutions in telecoms OSS/BSS

Billing system vendors used to build their own database systems, and network management system vendors used to create their own reporting tools. The reasons for doing this disappeared. Enterprise database technology got good enough to do the job at lower prices, and the reporting capabilities from specialist tools began to overtake those of NMSs. Consequently the OSS/BSS community stopped their own development and started building in commodity IT infrastructure and applications sourced from others.

Is the same process about to happen with data analytics platforms? After all, most of the IT solutions that telecoms operators and service providers want to invest in right now (customer experience management, proactive customer care, solutions to provide customer-specific and context-sensitive product offers) require ‘big’ data to be collected and analysed and acted upon (often in real-time or near real-time). And other functions, such as fraud management and revenue assurance (RA) can sit as an application on top of a data analytics engine.

Some niche analytics platform vendors saw this coming years ago. Lavastorm, for instance, has developed what looks to be a quite widely applicable data analytics platform on top of which many applications can be run – in telecoms, and in other industries. It has developed applications for fraud and spend analysis, revenue assurance and audits. Its telecoms customers include BT Global Services and Clearwire, Mobistar and Kabel Deutschland, TW Telecom and Telstra. The analytics platform can be used in conjunction with customer-developed or third party applications too.

The big IT companies are pushing analytics in a big way – Infosys’s BigDataEdge solution is a recent interesting idea. And these companies have been in the space for years.  In 2008 I wrote a report that predicted that the big IT companies with powerful data analytics platforms (such as IBM, HP, EMC and SAS) might reshape the market for telecoms revenue assurance solutions because they could replicate the functionality of the specialist players, and the investment case should have been much more compelling. I was wrong at that time: revenue assurance still exists as a distinct function within the telco, and the solutions used by that function are still in the main designed specifically to do a revenue assurance job (though many specialists are now owned by large, more consolidated BSS vendors).

The biggest of the specialist RA vendors, though, have tried to leverage the analytics capabilities of their systems to make them more widely applicable. For instance, Subex expanded the scope of its core fraud and revenue assurance products to become the ‘Revenue Operations Centre’, built around its ROCware analytics platform, and cVidya has built a series of revenue and risk applications on top of a common ‘ETL, security, reporting, and administration infrastructure’.

Why has revenue assurance survived, and why have other OSS/BSS vendors continued to develop their own data analytics platforms? Amdocs’s recently launched Proactive Customer Care solution, for instance, uses a data analytics engine developed in-house. I believe there are four reasons.

First, vendors naturally want to retain as much of the value of the solution for themselves as possible: using a third-party analytics engine, even if it is effective and sensibly priced, involves payment of royalties and involves some interface development effort (and ongoing maintenance effort) that vendors don’t want to spend. Vendors often say that ‘domain expertise’ is a prerequisite for building and using analytics engines, though they don’t explain why this is the case, or provide evidence that the providers of enterprise analytics solutions don’t have domain expertise (after all, the big players generally have very many more research, development and solution-builder staff than smaller, telecom-focused companies).

Second, in a way analogous to the fight to control the home gateway in consumer communications and media markets, there is a fight to control the data analytics layer in telecom IT. By using their own analytics platform, vendors have the ability to do more with the data, and one foot in the door for selling additional products that might use the same core platform; concede this part of the solution to a third party and it’s less likely the vendor will be able to sell more valuable, broader solution suites – buyers may opt for a ‘best-of-breed’ approach to OSS/BSS applications, increasing competition for the vendors.

Third, the ‘commodity’ enterprise analytics engines may be too expensive, or perhaps not up to the job required for some of the applications being developed. My view is that this may change as the collection of ‘big data’ in companies across multiple sectors drives development of systems designed to analyse it and expose it to other applications in a more efficient way. The rise of Hadoop-based databases in telecoms IT solutions shows that innovative technologies based on new architectures can displace well-established providers such as Oracle.

Finally, vendors – and their customers – simply may not be aware of the capabilities for commoditising the analytics layer. One of the things that constantly intrigues us at Innovation Observatory is the persistence of the silo mentality in telecoms despite 20 years of market pressures that ought to be removing it in the name of efficiency. It is still generally the case that customer experience management, CRM, revenue assurance or fraud management systems will all be specified and bought and run by different teams. The integration of IT and OSS in most telecom operators and service providers has not progressed very far in most cases.

And the relentless pressure to achieve improvement and efficiency in one’s own domain means there is often little time to think about whether approaches and technologies from other, even quite close-by, parts of the industry might be worth considering. For instance, in some ways customer experience management for mobile telecoms service delivery is still relatively immature compared with video quality of experience monitoring and measurement in the cable TV / IPTV market, where deep thinking about requirements has been carried out over a very long time, and sophisticated approaches and tools have been developed. Yet there’s almost no crossover between the specialists playing in these two markets.

So will OSS/ BSS vendors continue to develop their own analytics engines? Yes – for the time being. But I think in the next five years, the progress in the development of data analytics will mean that such an approach will be less and less sustainable.

Add this: