Cross Channel Optimization


Cross-channel optimization applies when more than one system touches the same customers. It is a general solution for all industries. Although it started in telecommunications, most users today are in healthcare and retail. In retail this optimizes store and online sales. In healthcare it improves health and reduces hospitalizations by optimizing clinical models delivered by telephonic nurses and pharmacists plus primary care physician.


The Nobi solution:
Nobi scientists developed a methodology to completely solve problems of this type called simultaneous statistical design. The mathematics took several years to develop and perfect in business use, but is easy and fastest for users. It can be delivered with no technical language or understanding by the client.


An Example From The Health Care Industry:





Building on statistical designs/analyses to evaluate ideally 20+ variables at the same time in a live testing environment, and taking our idea loosely from computer tomography, we realized that the same method could be used all at the same time across several channels. Since each channel tests 20+ ideas, which equals 1 million+ sales strategies built from the best of the 20+ tactics, multi-channel application brings enormous improvement power. Up to 5 channels x 20 interventions for each has been designed for. The mathematical power is hard to comprehend except through the results delivered. While the approach can be tweaked easily to follow changing trends, behaviors, markets and tastes, the main findings are sudden improvement, sustained long term.

In this way, performance is maximized by precise combinations of changes interwoven across channels.

Changes involving people, such as employees and customers, are also readily accommodated. The secret is 20+ innovative changes tested per channel, although strong results have been found with as few as 11 or even 7 per channel. The appearance that this will be complex does not carry.


A Retail Example:



The overall impression managers and executives tend to take away after one of these studies is that their own team’s expertise was fully harnessed but also amplified and “leapt over” to fresh innovations and profitable surprises. In-house experts know their craft but it can be hard to put the right finger on the customers’ pulse month after month, year over year without strong analytics.

Some Technical Points, Simplified:

An unfounded concern has been that the analyses may be contaminated by one channel interfering with another’s numbers and clouding real findings. In fact the mathematics keeps everything separate (and we have proven this theoretically and with simulations as well as with practical results for clients). The studies are multiply randomized as one of several simple devices that ensure pure results.

Another concern in prototype use of this method years ago was that it would be cleaner to analyze one channel at a time. In fact it is the other way about. The technical details are complex but it boils down to a similar feel to why we never put one rope on a sapling tree: usually we put three as each gives the others tension. In multi-channel sales, each statistical design in fact strengthens the others through the more strongly standardized methods that underlie.

The analysis uses standard statistical software adapted (by the analyst using a set procedure) for this purpose.