Analytical Engine

It appears that utilizing ATMs as a sales channel is not even a near future thing as we thought it was, it has already arrived.
Last Friday on my way home, I decided to withdraw some cash from an ATM of one of the larger private banks in India. When I was almost done with the transaction and was about to collect the money, there was a surprise waiting for me in the screen! The ATM offered me a reasonable amount of home loan and requested me to click a radio button written - if I want to avail or click a radio button with instruction of reminding me later if I want to think about it later or click - in case my answer was an outright no.
But then neither am I the only one, nor India the only country where it’s happening. One of our blog readers, a few continents away from my city, has already come across and availed one such offer from Chase!
Things surely are fast nowadays!

How many times one has deleted that email offering that free gold credit card or a pre-approved loan? How many times one has curtly disconnected the phone call from another tele-caller with a pre-approved check for that awesome holiday in Hawaii? In a market, where retail banks have exhausted almost all the channels to reach the target customers, ATM machines can be a powerful alternative for cross-selling.  ATM machines also provide the opportunity to reach out to the otherwise difficult to reach customers such as expats. 

So, how would it work? Almost every bank has pre-approved offers for certain customers.  The bank can load the offers for some of the pre-approved customers to the server that interacts with the ATM machines.  The moment the ATM recognizes a preapproved customer, it can show her the pre-approved offer through a blinker or an additional screen at the end of the transaction. If the customer is willing to go ahead with the offer, all she needs to do is to press “Y”.  The moment she presses “Y”, the bank gets notified and can track the lead through different mechanisms. For example, an SMS can go to the sales manager nearest to that ATM with the customer detail, who can take up the matter further.

Surely, it’s not as easy as it sounds. While the technology for such a process is available in the market, it’s still a challenge to implement it. Prediction accuracy is another big challenge.

·         It is not feasible to upload the entire set of pre-approved offers to the server.  And hence the need for a model, which can forecast with high accuracy the probability of a customer visiting a cash machine.

·         Each ATM screen on an average takes 5-7 seconds, which includes the time taken to interact with the server.  So, the time to pitch the offers must be chosen judiciously, especially for locations with long customer queues.  But restriction of timing will in turn affect the prediction accuracy.

·         The ATM cannot close the sell. The completion of the process depends on further integrations of channels.

Nevertheless, ATMs can still be used as an excellent channel and in the near future itself, with the advent of technology, it can be expected to be one of the primary channels.

�

Customers calling call centers don’t want long waits and often perceive them as longer than actual.

At the Intelligent Transportation Systems Lab of the University of Minnesota, an experiment was designed where highway driving conditions were simulated in a lab (a Saturn was peeled off and a 270 degree screen in front provided the video). It showed how people value time under different situations. While almost all the respondents remarked that they had actually ‘driven’ for a shorter time than the actual simulation, they also noted that they had waited on the ramp before entering the freeway for 30 seconds or more when the actual controlled waits ranged between 5 and 12 seconds.

A vital component of interaction between customers and service professionals is the time it takes to respond to the customer’s needs and requests or more importantly, the perception of such service times. Adaptive queuing theories can play an important role in improving perceptions of customer service. It is non-ideal to keep a customer waiting for a long time when their needs are fairly straightforward. And in such cases they value their waiting time even more than they actually waited for, in some cases, many times more. While such experiences may be fairly commonplace and their pain all too well-known, businesses, save for exceptions, rarely apply adaptive queuing theories to correctly predict and plan for servicing an individual. Disney engages customers in distractions while they wait for 45 minutes in a queue for a 3 minute ride. Situational elements like music, lighting, color, employee visibility, social interactions, visible queue movement, number of counters serving vs. idle, a visible clock nearby all influence perceived time. Companies can benefit by putting more thought into the psychological aspects of waiting times and queuing while trying to increase service levels and influence customer satisfaction.

One of the areas that our analytics center is working on is developing applications, which exploit the concepts of Behavioral Economics (BE), in a business context. We are looking specifically to develop a process to optimize a company’s online channel using BE concepts.

A recent WSJ (subscription required) article on Hewlett Packard mentions an interesting example which touches on two of the concepts we are exploring in our work.

1. Reducing the number of less relevant choices from websites
2. Reducing the uncertainty in a transaction process

For direct PC sales, where Dell remains dominant, Mr. Bradley reviewed customer survey data that showed H-P had too many PC models on its Web site, and reduced the number to 10 from 15. In January H-P cut the time it takes to reach a sales rep by phone to 22 seconds from 50 seconds. H-P increased its share of direct PC sales world-wide to 12.3% in the first quarter from 9.2% a year ago, according to Gartner.

We earlier posted about this topic here.

Customer risk segments are used to identify relevant product offerings. However, it seldom acts as a key service differentiator. Focusing on customer risk segments, especially as a grievance redressal tool, can help break down the century old cycle of (1) investigate, (2) verify, and (3) process the transaction. This three step procedure proclaims the customer guilty even before the verdict is out. Reengineering the resolution process can turn dispute resolution into an opportunity to build relationships.

When a customer disputes financial transactions, banks start internal investigations, verify the customer’s claims and then reimburse/waive-off the transaction amount or reverse the financial cost that has been levied on the customer. All these happen only if the customer claim is found to be true. But what about the grey areas where the verification process could be inconclusive?
Well, most retail banks have exception handling policies. Impressive!! Just that they are implemented long after the customer is fuming with anger. Is the customer going to be satisfied? The answer seems to be an emphatic No.
The three step resolution procedure can be strategically changed with the customer being “trusted” first, and the investigation, verification steps being done in the background. To implement this, we need a customer risk score available to all customer touchpoints (phonebanking, internet banking, etc.). The risk score identify customers that can be trusted.

Expertise in building  robust scoring models and process re-engineering can help identify process pain points that need correction. This can be leveraged to reinforce the relationship the customer shares with the bank, thereby increasing the lifetime value of a customer.

After porting my wireless number to one of the GSM cell-phone carriers, I called the customer care to activate my international roaming for my trip to India. It was a 40 minute long call, and the customer care agent explained that since I had been a customer for less than 90 days, he would need his manager’s approval to activate the service which would take up to 5 business days. When I landed at the Mumbai airport, the service had not been activated and 10 days later it is still the case.

The carrier not only lost potential revenue, incurred cost of a long care call but also caused customer dissatisfaction.

Is there any way for companies to avoid such one size fits all rules trap?

One potential way is to use predictive analytics to design the operational rules and automate front line decisions, to increase both revenue and satisfaction.

As jurisdictions across the US adopt electronic voting, the government administrations, vendors and users are climbing their respective learning curves in implementing new processes, developing new solutions and adopting new protocols at the polling locations respectively. As is almost always the case with large scale deployment of new technology solutions, there have been issues (see earlier related post).

Cook County in Illinois had its share of problems (go here for more details), where we recently had the opportunity to work with a team on determining the root causes for delays in transmission of election results. Given the significant implications of our findings to the players involved (the Cook County Board of Elections, the technology vendor), it was critical that the analysis be entirely data driven. To us the experience was very much like being part of a “Whodunit?” investigation.

We adopted a classic hypothesis driven approach to divide and conquer the problem with hypotheses ranging the gamut of possibilities across technology driven causes, process failures and governance gaps and break-downs.

Network applications and systems maintain rich data and that was the bedrock of bulk of our analysis. We could conduct rigorous and rich analysis of every component in the system that kept logs. The systems we dealt with, ranged from small embedded client devices, switches in the wireless network, servers in the central office to databases where results were tallied. The data sources were diverse, but the beauty was the utter simplicity of the analysis driven almost entirely by the clean structure adopted for the analysis - problems always seem simple to solve if divided into logical and small chunks.

The summary of our analysis was a time series plot of how cartridges moved through the system. This was a powerful visual, as it was entirely data driven. Yet, it made obvious the evolution of scenarios that caused a lot of stress on election night.

From a visualization perspective, we utilized a tool that enabled us to study correlations between the location of a voting device and the odds of successful transmission. We used a Microsoft tool called MapPoint and were thoroughly impressed. More details about it in another posting.

Note, the key word in the above question is ‘profitable’

Companies using multiple channels and/or multiple partners to sell their products usually track gross sales numbers to compare channel/partner performances. However, a top-line metric approach can be misleading. More importantly, getting to an accurate bottom-line view is not extremely challenging and does not require any significant IT investment or business intelligence tools.

Recently, we helped one of our clients to understand the profitability of their partnerships by taking a lifetime view of revenues and service costs of customers, acquired from the partner channel. The challenge we faced was in forecasting the attrition rate for each partner, to which our creative solution was to cluster similar partners together and calculate an average lifetime of each of the clusters.

In addition to lifetime revenues and service costs, we determined the cost of partner acquisition and maintenance (legal costs, business development and audit costs) along with the direct campaign related costs – to estimate a ‘true’ customer acquisition cost. An additional filter was added to control for variability of partner performance over multiple campaigns.

 

We discovered (to everyone’s surprise) that 86% of the client’s partners were unprofitable. More surprising was the fact that all these partners used to regularly pass the client finance department’s screening criteria as their metrics were focused on short-term benefits and not on ‘true’ lifetime value. Diamond’s analytical approach helped to fix the client’s issues and put the channel back on path to profitability (increased channel NPV by 190%)

It is interesting to note, that we did our entire analysis without requiring any support from client’s IT group. We directly queried client’s billing and other marketing databases to get the relevant data. Considering, how easy it is to perform the analysis; it is surprising that most companies do not have a good grasp of their partners in terms of profitability.

Our findings and more details of our methodology are documented in the whitepaper, Profitable Channels: The Right Metrics Make All the Difference

 

Expensive calls to customer care can be avoided by better managing core business processes. Most service providers are concerned with non-value added call volumes because they represent a significant cost to the business and offer little true opportunity for incremental sales. Often non-value added calls are triggered by problems or inefficiencies throughout complex operating environments, for example in claims processing. In our work we have found a stunning relationship between the “days to pay” intervals at several different types of insurance companies. As the chart below demonstrates, the probability of a claim generating a call to customer service increases dramatically as the claim ages. In this case claims that age beyond 40 days are basically guaranteed to generate at least one call, thus driving up costs needlessly.

So what? In this case, the company was tracking average claims “days of mail on hand” and believed that there were roughly four or five days of claims inventory to be processed. This faulty and misleading metric was well within the early part of the curve, where a call to care is highly unlikely. An accurate aging of the claims inventory revealed that the average payment interval was almost 17 days! This was well beyond historical norms and the operating assumptions that were used to forecast call volumes. By pinpointing the processing areas that had excessive aged inventories (using further detailed analysis of electronic data) we were able to devise a simple plan to reduce pockets of aged claims, put more accurate measures in place, and stanch call volumes.

Why most executives’ are disappointed with the results of a new process rollout?

Because, when companies replace an existing process and associated technology with a new one, the realized benefits (additional lift in number of registrations or customer approvals, gains in productivity etc.), in most cases, is way short of the projected benefits.

Having worked in a few such engagements across industries, we know that to get to the root of such problems, it is very important to form good hypotheses around what has changed (process architecture) and how is it being measured (accuracy measurement).

1. Process Architecture: New system architecture is not identical to the old one, and neither is the baseline performance of the new process
2. Accuracy Measurement : Either due to change in process architecture or statistically inaccurate measurement, there are biases, which have not been accounted for, resulting in unexpected results

This is followed by isolating the data to test each of the hypotheses individually. In many cases we end up performing fairly powerful and sophisticated statistical analysis with system log data and historical transaction data, the kind of data one normally does not associate with analytics techniques.

Normally the first step in such projects is for us to de-aggregate the process and its supporting technology and build what we call a component interaction diagram which is an end to end mapping of the associated systems – which forms our baseline understanding of the process. Data is collected for the relevant systems and transactions and analyzed to prove or disprove the hypothesis, one at a time - till we isolate the reason or set of reasons explaining the performance gap of the process and identify how they can be fixed to get the promised benefits.