When Will Customer Analytics Tools Deliver?
A version of this article first appeared in eB-21, published 10 times a year in Europe by TBC Research .Based in London and San Francisco, TBC Research helps senior business professionals make more informed technology decisions through its magazine, research, and events portfolio.
AMR analyst Kevin Lucas is puzzled. Research about to
be released by his firm shows that customer analytics projects are among the longest and most arduous to implement, but when completed provide some of the best returns around. The rewards are obvious--maximising revenues from your best customers will bring a smile to the face of most CEOs--but the reason for the implementation difficulty is not quite so straightforward.
The answer to the problem lies somewhere between the uncertainty of the vendors, the immaturity of the market and the confusion around what people are trying to achieve. At its most basic, customer analytics is designed to tell you who your most profitable customers are. But beyond that simple calculation of revenue minus cost-to-serve there lies a whole new, largely unexplored world.
In fact, although analytics was the second most beneficial of the 12 CRM project types AMR researched, it still only managed to meet, rather than exceed, expectations. Lucas explains this as a failure of the figures to hone in on things like measuring customer profitability--where extra revenue is generated over time and is geared towards the long term--but admits to being worried by the underachievement.
"We're not seeing what the vendors are promoting," says Lucas. For example, although many vendors are promising implementations of 90-days or less, analytics takes on average over 11 months, one of the longest project installations. In addition, the chances of success are evens, which means many are less than successful. Lucas will not commit on whether this is because of difficulties in getting the data together in the first place, itself a major task, or because the products are difficult to configure. Much of the software is so new it's difficult to say.
The immaturity is both symptom and a cause. Packaged software vendors traditionally refine their offerings based on the experiences of their early adopter customers, but when users are not sure about what they are trying to achieve from a project, packaged software is less useful. That's why some of the early suppliers are making a big play of the intellectual property bound up in their products.
Charles Grover, director of CRM for EMEA at PeopleSoft, which has tied up with Balanced Scorecard inventors Kaplan and Norton among others, says, "The difference between us and most other offerings is we prepackage the intellectual capital to say what the end result should be, rather than saying 'there's the technology now go and analyse customer profitability.'" PeopleSoft offers a layer of behavioural scorecards, describing things like retention and profitability.
Nortel, meanwhile, has trademarked the term "return on relationship" and aims to get people started with 44
out of the box reports for sales, support and logistics--including reports like campaign effectiveness--hit rate ratio, call volume velocity and change request metrics. Paige Mantel, head of product marketing, admits these will not cater for everything but adds: "It helps customers get up and running with the reports they might want to use immediately."
With much frantic positioning in the market, prospective users looking for a customer analytics product can be easily confused. The big operational suppliers have all recently launched comprehensive customer analytics suites, no doubt in response to warnings from analysts like AMR that there was too much emphasis on operational CRM and little, if any, analytical content in their offerings. Last summer, Nortel Networks, parent of Clarify, introduced a full suite of CRM analytics through a partnership with Broadbase. In December, PeopleSoft, supplier of the Vantive products, weighed in with its offering, building on its performance management platform ePM. It has many point customers, but struggles to point to a full analytics suite use. Siebel, the gorilla of the operational market, recently partnered with SatMetrix for its offering.
The traditional analytics vendors have also been shifting their focus to the customer. Brio, Cognos and Business Objects all have CRM offerings, the latter setting up a
dedicated division called Ithena. While much of their attention is focused on historical profitability analysis, at the high-end, heavy data mining vendors like SAS Institute provide products which are more predictive. For example, SAS claims to be able to tell when a customer is likely to defect and set up the appropriate response. "Traditional analytics can only tell you what has happened," says Peter Dorrington, business solutions manager at SAS. "It's completely retrospective. Data mining starts with the question, then moves on. It reveals hidden patterns because it's good at dealing with the detail. You need a range of tools and techniques."
The future, meanwhile, is about real-time, automated analytics. Taking customer data off-line, into a data warehouse or similar data store, is where most of the focus is today. But by the time an 11-month project has discovered who the company's most profitable customers are and returned that data to the business, those customers' circumstances may have changed. "People think real-time CRM and off-line analytics are completely separate," Dorrington says. "The reality is you need a marriage of the two."
In February, Giga produced a report entitled "CRM analytics: what's missing?" It stated:
"Instead of just tracking the numbers and spitting out reports, the applications should be driving business processes based on these analytics. All current offerings stop short of this, delivering only a partial solution and ignoring the most interesting part. Vendors should be providing the necessary hooks in their routing engines, scheduling systems, pricing catalogues and notification methods to allow events to be triggered by these metrics. If they have not architected their systems to allow this today, and are not in the process of delivering 'out-of-box' automations in the future, they have completely missed the point of the exercise."
But if that sounds simple, there are significant challenges around providing real-time automations. Simon Jennings, VP of Northern Europe at data extraction vendor ETI, says: "One of the first issues when you put an architecture together for CRM is are you doing it in real time--in which case you need some messaging middleware--or are you going to do some data caching? Normally the amount of data needed means you will do caching."
In an effort to bridge the gap between operational and analytical CRM, Compaq is currently engaged in an initiative it calls the Zero Latency Enterprise (ZLE). Targeting the specific verticals of telecoms, financial services and retail, it is putting together applications, platforms and databases so that information about a transaction becomes available across an enterprise the moment it occurs.
Whatever the future holds, if real-time and writeback do eventually arrive, the goal will probably move on again. Already, academics are starting to define a new area that could become the next holy grail in customer analytics--measuring customer