Real-Time Decision Managers can’t do EDM but they are useful for Service

I was asked by a Bank recently if we had an ‘Event engine’ that could handle and process 200,000 ‘Events’ per second.

It seems that they are interested in Event Driven Marketing and we have had some discussion about this, but they are receiving advice from a leading consultant, who told them this would be a minimum requirement for implementing Event Driven Marketing in their bank.

I got to thinking about this and I have to disagree with my esteemed colleagues at the consultants.

Why? Well, my guess is that they think Real-Time Decision Management (RTDM) engines are the same as Event Driven Marketing. They are not.

Research shows that the number of Significant Events happening to any particular Banking Customer is 1.4 every 5 years (Source: Banking Strategies). This works out at 0.28 Significant Events per year – or about 280,000 Events per million customers. This in turn gives almost 1,000 leads per million customers per day.

This research is backed up by the reality achieved by our customers in Italy, Hungary, Austria and the UK along with the published results from Australia and Norway, who each generate about 700-1,000 leads per million customers per day.

EDM_fishingLet’s say there are around 1,000 actual Significant Events per day compared to the 200,000 per second (17,280,000,000) that need to be monitored according to our ‘consultant’. Alternatively this is 280,000 Significant Events per year compared to the 63,072,000,000,000 (63 Trillion) being monitored by the real-time system.

That is 1 Significant Event for every 17.2 million interactions.

Of course the reason for this discrepancy is to do with the definition and understanding of what is Event Driven Marketing.

Let’s start with what is an Event. From the Wikipedia page we learn that “A Significant Event is a major happening in a customer’s life. They can lead to a measurable change in a customer’s normal behaviour, state of mind, personal circumstance, or interaction pattern. It offers a reason to communicate with the customer, with a relevant proposal, at the right time.”  It is interesting to note that one major Dutch bank presented that 55% of all their product sales came from 5 life time Events.

Here are some examples:
•    Receiving a windfall.
•    Starting education.
•    Starting work.
•    Moving house.
•    Getting married.
•    Having a child.
•    Stopping work.
•    Starting a business.
•    Becoming unemployed.
•    Retiring.
•    Leaving the bank.

Any one of these Events can lead to major changes in a customer’s financial needs, and this is the opportunity for the Bank.

Interestingly, almost none of these can be detected by RTDM or in Real-Time and even those few that can be detected in this way there is no desperate need to provide a response in real-time.

So RTDM can’t find most Significant Events. Well what can you detect with RTDM?

Here are some examples:
•    Customer is using our services at the call centre/branch/ATM/Kiosk/internet.
•    Customer has logged in.
•    Customer has landed on web page.
•    Customer is within 50m of the Branch.
•    Customer has deposited/withdrawn/transferred some money.
•    Account is overdrawn
•    Account’s balance is </> x.

Events_timingThe graphic (right) shows a these Events and maps where they can be detected, their relative value and the period of time required for a response (seconds, hours, a day, a week). The fact is most Events remain valid for days or even a week. Do you need to contact a customer when they have a baby within 1 second of the birth or will this still be valid in a week’s time? On the other hand what is the value of knowing that a Customer’s account was overdrawn last Tuesday?

So what is the objective of such RTDM systems? Well they detect Customer interactions with the Bank in real-time and then ask the question “What do you want me to do?” Well, looking at several vendor sources you come with one main and one minor answer. The main objective can be typified by the statements below:

  1. If the data isn’t evaluated fast enough, and the offer isn’t served up immediately, it doesn’t matter if it’s the most relevant offer in the world; you’re likely going to lose that opportunity with the customer,” says Todd Denton, VP/GM, Financial Services, Epsilon.
  2. The concept of real-time marketing is not new—for decades, brands have used CRM solutions to organize customer information and tee up the “next best offer” to be communicated in “pull”/inbound channels, such as call centres or branches.
  3. Get maximum value from your inbound, real-time customer interactions. Track and respond to customers at every possible touch point – service, sales and support, etc. to deliver the best real-time recommendations and decisions to your interactive customer channels – websites, contact centres, and point of sale. SAS
  4. The aspiration of most financial services marketers is to leverage the most recent consumer insight available to deliver offers that are both timely and relevant. Greater access to consumer behavior online, and the ability to process and convert this insight into dynamically personalized content and offers, is helping to make ‘real-time marketing’ a CRM reality for the more advanced banks and credit unions. Financial Brand 
  5. Real-time means driving more cart and content usage – drive recommendations in real-time to get more content read, more purchases in the cart (of course this one is a retail example).

Ah, so we see that the main objective of the new RTDM system, as stated by most vendors, is to make timely and relevant offers to your customers, in other words this is a real-time Product Push mechanism.

This may well work in the retail sector where customers make product purchases frequently. For example, see number 4 above and read it in the context of a retailer and it makes much more sense. I can see why someone like Amazon would look at my online behaviour and convert this into dynamic personalised content in real-time whilst I am on their website, but I fail to see in what way a Bank would do this.

As you will see if you do your research, the vast majority of all use cases for RTDM use retail examples. There are almost no financial examples (and those that exist are poor). In fact, vendors (and consultants) are still struggling to find that magic scenario.

So, as we have seen RTDM (in banking)

  • Can’t detect Significant Events and
  • There is only 1 Event for every 17 million customer interactions.

If you dig a bit deeper you can find the minor objective for using RTDM from a couple of more advanced vendors which can be characterised as follows:

  1. Improving the customer experience and driving customer engagement continue to be high priorities for many organisations
  2. Real-time can mean cross-functional customer experience management, next best action not just next best offer, bringing service and other actions into the picture beyond just marketing.

That sounds a lot better and makes more sense (to me).

In fact I think this is the crux of the problem. Vendors have developed RTDM systems and have found some level of success with them in retail and they think that this will transfer directly into the financial sector. But retail is predominantly product sales focused, whereas finance is service focused. Just think about it, how many products do you sell and how many times are they used? For example a credit card gets sold to a customer once, but it is then used 200-350 times a year for the next 5 years. That’s one sale compared to more than a 1,000 uses – each of which is a service. Thus vendors need to look at the service industry to find examples of how to use RTDM and the above two comments are some of the very few that I have managed to find.

In summary, RTDM systems don’t have the capability to deliver Event Driven Marketing for Significant Events nor do I believe that real-time product push (in banking) will have any better results than previous efforts. So when your vendor tells you that this is the way forward, I suggest that you challenge this.

Where I believe that RTDM can fulfill a vital role in your marketing is by providing service to your customers, enhancing their experience, increasing their satisfaction and loyalty to you. For this you would need to be able to monitor every kind of interaction and make recommendations. Wait a minute, for that you would need an engine that could cater for 200,000 interactions a second. I wonder where I could get one of those………….

For those of you interested in detecting Significant Events in your Customer’s lives, come and talk to us. For those of you who are interested in enhancing your Customer’s experience in real-time, go see your local RTDM vendor – but don’t expect to get EDM, even if they say so.

facebooktwittergoogle_plusredditpinterestlinkedintumblrmailfacebooktwittergoogle_plusredditpinterestlinkedintumblrmail
twitterpinterestlinkedinrsstwitterpinterestlinkedinrss

2 comments to Real-Time Decision Managers can’t do EDM but they are useful for Service

  • Laszlo Rekasi

    Hi Mark,

    Thank you for sharing your thoughts! Very interesting what you write.
    However I can imagine a mixed world. The RTDM can be used in branches where we, as a bank recognise the reason why the customer comes. Meantime we have an offer/story-pool (basically the same product/service is packed with different, personalised marketing message). Based on the reason of the interaction we may have a RTDM to select the proper offer from the pool or even generate a new one.

    On the other hand if you want to set up a multichannel offering environment then you need to scale. This system which is very close to RTDM. Processing logs during the night takes time and if anaything goes wrong you have to rerun it. Sometimes you have no time to process such amount of data if faild last night. Scheduling the dataprocessing in more frequent timeframe can solve this issue. it is up to you how frequently you want to process i.e. transactions, netbank logs, website interactions, etc. There are tools which can manage such load mentined in your article. You can find many in NOSQL world. These scale-out easily and are cheaper than traditional RDBMS like Oracle. Of course, you have to use the right tool for the given problem.

    Best regards,
    Laszlo Rekasi

  • Hi Laszlo,

    Thanks for your comments and feedback. It is much appreciated.

    I too think there will be a mixed world. This was my point. It is just that RTDM is not good/able to recognise significant Events, but it is VERY good at providing service, etc. in real-time. So, as you say, use the right tool for the problem.

    My concern is that RTDM is being positioned by vendors specifically as EDM. One vendor is currently talking to Erste saying they can do this, when it becomes clear that they don’t even understand what EDM is. The same is happening with another vendor and a bank in Poland.

    The reality, as I tried to show in the graphic, is that both are complementary to each other and can co-exist happily – but other vendors don’t want to co-exist. They want complete control of the customer.

    As for Next Best Offer scenarios, I actually plan to talk about this in my next blog, so stay tuned and then you can give me feedback on that :-)

    Cheers,

    Mark