The De Villiers Walton Blog
Welcome to the De Villiers Walton blog.
SAP RTOM 7.1 - Promise vs Reality
SAP RTOM 7.1 - Promise vs Reality
What web channel manager hasn’t used Amazon’s recommendation engine and thought, “I want one of those”?
Well, imagine a software solution that – like Amazon – presents recommendations and offers to the customer in real-time via the web.
Imagine that the software is able to “learn” from a customer’s response and is able to make recommendations and offers based upon what it has learned about that customer.
Imagine that the solution works not only via the web but also via call centres and, in the not too distant future, mobile devices.
SAP Real Time Offer Management (RTOM) promises all of this today with SAP RTOM 7.1.
In this piece, we assess to what extent the promise of RTOM matches reality.
At De Villiers Walton, we’ve been following the development of RTOM since SAP’s acquisition of Ingeneo in 2007.
While excited by RTOM’s potential from the start, SAP’s first attempt to integrate RTOM within SAP CRM 2006s did not convince us that the product was ready to take to our customers. Despite this, we continued to follow the evolution of the product and in January 2012, a team of CRM and BI specialists from De Villiers Walton spent a week with the RTOM development team at SAP Labs in Israel working with and learning about SAP RTOM 7.1.
So, what is RTOM?
SAP defines RTOM as a, “self-learning multichannel real-time recommendation engine that recommends contextual optimal product and next best actions, that are likely to be accepted and provide the desired business value”.
Let’s break this definition down a little.
1. RTOM is a real-time recommendation engine
Real-time, real-time, real-time.... A key advantage of RTOM over functionality offered elsewhere in SAP CRM is the fact that real-time offers and recommendations are made either by reacting to how a customer navigates around a website or by reacting to the responses given by a customer when speaking to a call centre agent.
The offers are determined in real-time and are derived from any number of data we have for a particular customer including master data, transactional data and customer responses to previous offers and recommendations. With RTOM’s learning capabilities, where it captures and analyses customer response to recommendations and offers, the relevancy of recommendations and offers to a particular customer increases over time.
Let’s consider two typical customer interaction scenarios – in-bound telesales and website sales and see how RTOM can help both SAP and non-SAP customers to take advantages of its unique and innovative features.
Inbound telesales interactions where a customer calls the enterprise to order a product or service, request assistance, pay a bill, make a complaint or enquire about a product or service are far less intrusive than outbound telesales interactions.
An inbound call presents a brief window of opportunity for a call centre agent to try to cross-sell/up-sell the customer, retain the customer with an improved deal, capture a payment from a customer, update customer contact details and so on.
At the start of the interaction, initial offers and recommendations use data we have about the customer before the interaction commences. Then, as the interaction proceeds, additional offers and recommendations are made in real-time and can be based not only on the data we had about the customer at the start of the interaction but also on data captured within the current interaction including the customer’s response to other offers and recommendations.
If, on the other hand, the customer receives an inappropriate offer or recommendation by for example, recommending a product the customer has previously purchased or rejected an offer for; the outcome can be quite negative. RTOM addresses this issue.
RTOM works in a similar way within the web channel. Offers and recommendations made at the beginning of the customer’s interaction with the web channel can change according to how the customer navigates around the website. While at first glance, RTOM provides a customer recommendation experience similar to that offered by Amazon in reality the level of responsiveness can be much greater.
2. RTOM works across multiple channels
Currently RTOM is available for the web and call centre / interaction centre channels. SAP Labs gave us a taste of the future with an impressive demonstration of RTOM on a mobile device within a retail environment. We look forward to this functionality being available in a future RTOM release.
By working across multiple channels, RTOM allows us to either differentiate the offers and recommendations made to a customer according to the channel or by integrating these offers across channels. For example, if a customer has already accepted or rejected an offer through the web channel does it make sense to make the same offer to them when they speak to a call centre agent?
Of the two currently supported channels, only the interaction centre is integrated fully into SAP CRM i.e. offers can be defined in SAP CRM Marketing and then presented in real-time to call centre agents within the Interaction Centre.
RTOM is not integrated with SAP CRM Internet Sales or Web Channel Experience Management (WCEM). Any web integration requires bespoke development to enable recommendations and offers to be presented in real-time to the user. An SAP customer has two choices when implementing SAP RTOM for the web:
- Implement RTOM as a standalone solution using the RTOM toolkit to define customer offers and bespoke develop the web channel to show the recommendations.
- Implement RTOM within SAP CRM Marketing to define marketing campaigns with customer offers and bespoke develop the web channel to show the recommendations.
3. RTOM is self-learning and recommends products and next best actions that are likely to be accepted.
This is one of the most exciting aspects of RTOM but also the most troublesome from an implementation perspective as it requires a slightly different approach to solution design, build and testing.
We believe that proofs of concepts are going to be critical when selling RTOM solutions to customers. Because RTOM is effectively a “black box” with its own proprietary algorithms, understanding the system outputs from any number of inputs is difficult to say the least.
While it is possible to define recommendations and offers and “ignore” the learning, the true value from the RTOM solution comes from not only the real-time delivery of offers but the system’s ability to learn from responses.
We believe that proofs of concepts are necessary in order for the customer to gain confidence in RTOM’s capability.
Let’s revisit that SAP definition and boil it down to the basics – RTOM is a self-learning multichannel real-time recommendation engine that seeks to make the right offer or recommendation to the right person at the right time.
The ideal customer
The ideal RTOM customers are medium size or large enterprises in the financial services, utilities, telecommunications, retail, consumer packaged goods or high-tech industries.
These customers will not be “green-field” sites. They will be established users of SAP CRM and SAP BW.
They will use RTOM to deal with business challenges concerning customer loyalty and retention, maximising customer profitability, effectively selling a complex offering of products and services, managing a rich pipeline of new product launches or the need to interact with customers across multiple channels.
From our perspective, the ideal RTOM customer will typically also be an early adopter of new technologies.
RTOM 7.1 is the latest version of RTOM and offers out of the box integration with CRM 7.1 for Interaction Centre and comes with standard business content for SAP BW.
There are currently two SAP CRM-based industry-specific solutions available – Telecommunications and Utilities – along with a generic cross-industry solution.
RTOM can also be installed standalone without the need for any other SAP components.
RTOM has two key components – the Offer and customer Experience repository. Users design and build Offers in one of three different ways – through the RTOM Business Logic Studio (BLS), through the CRM Offer Designer, embedded within the CRM Marketing Campaign or through third party or bespoke tools that are capable of compiling offers in the RTOM XML format.
The Offer contains information about the customer, agent profile and the product or service that is the basis of the offer. The Experience repository contains the customer feedback (Accept, Reject etc) about the offer. RTOM’s self-learning engine utilises the experience data and makes offers and recommendations that are more accurate as the customer experience builds up over time.
Eyes wide open
Before embarking on an RTOM implementation, we believe that you should consider the following:
- Currently, the RTOM self-learning engine is based on a proprietary learning algorithm and does not support standard algorithms such as decision tree, neural networks and association rule learning etc. We believe that SAP would not utilise the RTOM algorithm without close scrutiny. However, without a deeper understanding of how this “black box” works in comparison with other well established machine self learning algorithms, some organisations may be reluctant to consider RTOM.
- Currently, RTOM’s out of the box integration with SAP CRM is limited to the Interaction Centre only. If a customer has SAP CRM Mobile, Internet Sales or Web Channel Experience Management (WCEM), they will need to perform a custom development.
- Only standalone mode supports clustering (i.e. more than one instance of RTOM). Out of the box, RTOM only works in a single instance mode with the SAP CRM Interaction Centre. This creates unnecessary complexity in an environment in which a customer has an SAP CRM Interaction Centre and a website and plans to run RTOM for both channels.
- The RTOM Business Logic Studio (BLS) and CRM Marketing campaign based offer designer provides two completely different user interfaces and offers different levels of usability. BLS has a textual based UI and requires a deeper understanding of the RTOM expression language that seems to be a combination of SQL and regular expressions. CRM marketing campaign based offer designer has a graphical UI and is generally more user-friendly. However, the CRM UI doesn’t have the auto-complete functionality of BLS when a user is writing the business rule using the RTOM expression language. In our opinion, this is an omission from SAP.
Is RTOM the perfect product? No, of course not, no new software product ever is. However, if you fit the ideal customer profile, believe in the power and value of making personalized offers and recommendations to your customers in real-time and want to gain an advantage over your competitors, it warrants your serious consideration.
If RTOM is of interest to you we strongly recommend that you trial the technology via a proof of concept or small-scale pilot before embarking on a full-scale implementation.