The “furor” over artificial intelligence and especially the “ChatGPT” technology is well-deserved. On the one hand, the idea of AI-generated communications, instructions and even critical decision-making is a natural evolution in the advancement of technology. On the other hand, there is intelligent, well-respected schools of thought that it’s simply too much too soon.
By: Rob Hand
CEO, Hand Promotion Management, LLC.
April 3, 2023
Anyone not living under a rock these days is somewhat well-versed in the acronym-based language of text, Twitter® and social media. One of my all-time favorites is LoL, or what is commonly known as “Laugh out Loud.”
Another acronym I see a lot is “IDK”—short for “I Don’t Know.”
Recently, I’ve heard and seen that expressed verbally and acronymically about artificial intelligence and machine learning – AI and ML.
In a recent text with the CIO for a major pet foods brand, I asked, “Where do you folks stand with a decision on promotion optimization technology?” Her response: “IDK.” Likewise, when I asked about potential TPO vendors of their VP of sales ops who was in charge of trade promotion management, I also received an “IDK – maybe nowhere” reply.
I have blogged, podcasted, gave speeches and held workshops on TPx technology, and invariably, the one area of indecision is what to do about TPO. One of my roles typically performed in pre-TPx transformation initiatives is to discover, assess, and analyze the requirements across the various internal organizations to be impacted by trade promotion in general and promotion optimization in particular.
My viewpoint (Yes, OK, “POV”) here is from the perspectives of the end user and the consumer. I don’t profess to be a data scientist or Python developer, but I’ve designed enough scenarios, workflow and UX for TPOs to understand what it takes to build a successful solution. I also work with so many of the data providers to know what source data is required to generate trustworthy and precise predictions.
Everyone wants to consider TPO as an eventuality—heck, even a mandate in the transformation of end-to-end promotion planning, execution and analysis. Just not right now.
From that experience, it is easy to see how TPO gets sidelined. For some specific reasons why there is a reluctance to jump into the onboarding (or development) of TPO, check out my previous blog on this HERE.
Of the top commercial TPx software solutions, many of them have TPO functionality. Most of them have TPO functionality built into the planning architecture and some of them have stand-alone solution modules. Even some of those with the embedded TPO functionality partner with analytics vendors which have more powerful optimization engines and predictive technology.
However, the rush to get TPO technology onboard is premature because, as we all know, the lack of good data degrades and/or prohibits the accuracy and trust of most outcomes. Sure, there is typically some sort of “best case” result, but how much a KAM is willing to stake the fate of the deal is highly questionable. If I ask about the feelings of confidence he/she has in science that produces the prediction, I most assuredly will get “I don’t know.”
During the research for the latest survey on trade promotion, we asked many questions about TPO plans, capabilities, usage and satisfaction. For the complete TPO survey data, request your copy of the HPM 2022 Survey on Trade Promotion. Part of the survey we did not publish were the hundreds of comments we received—many of which were about TPO. Here are just a few:
- “Does the TPO work for us? I really do not know.” – VP Sales
- “My personal assessment is…I don’t know.” – Director of Trade Operations
- “I don’t know if our KAMs use it or not.” – Senior Director, Demand Planning
- “You may need to poll each of them, but as the VP Financial Operations, I don’t really have a clue!” (IDK)
- “I do not know if I would show my buyer the basis of the promo plan that was optimized on our [TPM] system—probably not.” – National KAM
- “I don’t know if the basis for the proposed [optimized] plan is sound or not.” - CIO
- “I don’t know, maybe it is, but I don’t know what the science is.” – EVP Sales
- “My problem is that I don’t know what I don’t know!” – Division CFO
Lots of IDK’s there.
Here’s the bottom line: We are pushing hard to develop, deploy and execute trade and channel promotion optimization using sophisticated and advanced AI and machine learning. The good news is that there ARE some very powerful and well-developed solutions out there that have excellent data science at the core. There is also a growing corps of data scientists who are becoming extremely adept at consumer goods industry domains and who know how to identify, model and write good algorithms that produce very believable and USABLE predictive promotional plans.
The frustration felt by promotion planners is real. The initial use of a TPO is going to feel very good, and perhaps it is because it seems very cool to press a button and have a solid set of answers that you would have had to work hours to get to, doesn’t it?
But remember, the most common answer given by sales reps and KAMs to the question of purpose for trade promotion funding is to provide financial incentive to get the sell-in deal. So, how valid is their assessment of the precision and trust of the optimized plan? Do they really even focus on the end objective of the promotion at the shelf or how best to get the volume?
Maybe that charge is a bit harsh, of course. In many cases, it is highly likely that the eventual tactics, timing, and even promoted product group will be changed by the buyer before a signed deal commitment.
That said, once the post mortem is done on the promotion, there is a near 50% chance that, after all that wrangling of promotion optimization scenarios, the promotion still fails to attain the level of ROI expected. The best reps and KAMs actually do run these optimization scenarios and work hard to create a strong promotion. But when the data is bad and the AI/ML engine fails due to lack of granularity, the planner may not even know that.
The numbers do not favor trust. 48.7% of the more than 300 CPG executives polled do not believe TPOs produce trusted, accurate predictions of outcomes. 41.2% do have a favorable view of TPO accuracy (Based on the results from the HPM 2022 Survey on Trade Promotion).
Oh, and what is the response from the other 10.1%?
Yep, “I don’t know.”
Artificial intelligence is being hailed as significant human achievement while, at the same time, projected as a scary, fearful potential for robotic manipulation of human frailties. Twitter®, Facebook®, Instagram™, YouTube™, and virtually every other social media platform show everything from AI/robotic utopian paradise to something just north of the zombie apocalypse!
For trade promotion, revenue growth management, retail execution, and integrated business planning, there is no doubt that AI and machine learning hold the promise of smarter, more effective promotions, higher success rates for new product introduction, more resilient supply chains, and higher quality consumer engagement and experience. There is indeed a lot of great work being done in the arena of artificial intelligence, and patience is clearly called for.
However, that said, promotion optimization has yet to reach maturity, hence unanimous trust among the key players in the promotion planning and execution drama is a long way from happening. In addition to the continued improvement of advanced AI/ML technology and data, for universal acceptance of TPO to happen, something more is needed—human buy-in and support.
Even as the TPO technology improves, another sea change must occur. The relevance it has as a mission-critical requirement for promotion planning depends upon the consistency of accurate and dependable causal factor-driven outcomes and the trust everyone in the demand chain has the TPO results.
That is only going to occur with long term consistent performance and a VERY minimum degree of error between forecast/planned and actual success of the promotion. To get there, more effective and frequent collaboration is going to have to expand not only between corporate sales and marketing, but with the channel partner as well. This is something that we have promoted for years, decades really.
Aside from the obvious need to drive better, cleaner, more harmonized and aligned data into the TPO functional framework, there is a tremendous amount of enrichment of understanding among the major stakeholders and executives on both sides of the trade channels before we can declare promotion optimization a mature science. We can see the progress, even though the frustration and doubt still permeate the corporate infrastructure of the demand chain. But I am confident that we will see it happen sooner than many will believe.
How long is it going to take to get to that point?