JTPP: The Real Issue Behind Promotion Optimization
By Kevin Sterneckert, research director, retail
at AMR Research
Through a special arrangement,
presented here for discussion is an excerpt of a current article from The
Journal of Trading Partner Practices (JTPP), the official online publication
of the Vendor Compliance Federation (VCF), Trade Promotion Management Associates
(TPMA), and the Federation of Credit and Financial Professionals (FCFP).
With a typical food retailer
realizing up to 60 percent of its revenue from promoted items, why are
there considerably more price optimization stories than promotional ones?
AMR Research’s studies suggest there are less than 10 retailers engaged
in any form of promotion optimization. Remember, it’s not just about the
ability to optimize promotional pricing, but the size and complexity of
the business process as well.
Despite the quickly maturing
science with its proven success, today’s applications reinforce the broken,
linear process, which isn’t the reality of how retailers promote their
assortments (see Figure 1). These typical five steps to creating a circular
across hundreds or thousands of items and flavors create an unwieldy and
sub-optimal workflow for merchants.
Successful promotional optimization
stimulates consumer response by finding the right price for the right shopper
segment in the right store, requiring the appropriate blend of science
and art conducted in a time efficient manner.
For a typical food retailer
that promotes up to 3,000 items per week, scalability becomes an issue,
with item level decisions next to impossible. Delegate these activities
to analysts, and the merchant becomes removed from the decisions of what
merchandise to promote.
AMR Research suggests letting
the science and business processes handle the initial administrative work
of creating suggested promotional candidates for a given promotion week.
This new approach removes administrative steps from the merchant’s desk,
allowing the merchant to be a merchant. In the art-science equation, the
science virtually eliminates and/or consolidates significant activities
into a single step to deliver the candidates that meet the objectives of
a given promotion week (see Figure 2).
Be sure your application
provider’s functionality includes the following capabilities. Doing so
will deliver the right conditions to succeed in implementing promotion
optimization across the enterprise:
- Start where you end today, recommended
promotional items. Let the software and science do the heavy lifting
and simplify the entire process. - Include robust workflow/task management
as well as role-based activity management and reporting. - Demystify the black box (science). Merchants
should be able to understand the reasons behind a recommendation. - Don’t forget versioning optimization.
True consumer-centric merchandising suggests that circulars should differ
(items and prices by location or ad zone). - Integrate or include ad planning tools.
- Have support for trade fund management
(not manual entry). - Include total business halo and cannibalization
impacts. - Embrace customer segmentation capabilities
and loyalty pricing. - Be sure to have e-commerce capabilities
for cross-channel retail. - Ensure the single demand engine receives
signals and sends to the supply chain.
Discussion Question: Why aren’t
promotion optimization tools being utilized by retailers? Are the inherent
conflicts between the art and science side of merchandising bigger around
promotions than pricing? What do you think of suggestions made in the
article for implementing promotion optimization?
After reading this article, I’m more confused than ever. Isn’t promotion optimization when you take a hit on a product to get people to buy full-margin products? That view is a lot simpler than what is presented here. And a simpler term for it is shielding. This stuff is too complicated for field operators. If shampoo is on sale, I’m clip-stripping brushes and combs on the display. That’s how I look at promotion optimization.
Promotion optimization makes a lot of sense, if it can be easily executed; and that is where the problems begin. Retailer chains are not set up to do this. Yes, you might see a few items priced differently by store location, but really diving into the numbers and coming up with innovative solutions remains out of the grasp of most retailers. Over the years, retailers have passed much of their analytical work to manufacturers. Few still have the wherewithal to undertake a full-blown promotion optimization program.
Do these programs make sense? Yes. Will we see many chains adopt them? Probably not.
To me, the real issue behind promotion optimization is not art or science, but data. Customer data in particular. It’s interesting that there may end up being a debate in this column as to what promotion optimization even is, because to me it is taking the principles of price optimization and applying them to a subset of customer purchase history (usually based on segmentation) to identify price point differences from “general” demand, and use those price differences as promotional offers. You can apply it at the general level of population, but it’s more effective when you apply it to segments.
With that definition, the challenge is in getting clean customer purchase history, and having agreed upon customer segments that both marketing (who communicates the offer) and merchandising (who shapes the offer) use and understand well. I’ve seen many a company never make it past the segmentation issue, let alone scrub their customer purchase history data to the point where it can be an input into a sensitive price optimization algorithm.
Lots of retailers that I’ve seen attempt promo optimization have found themselves in data quality and master data management initiatives instead–when we work through those challenges, then I’ll be interested to see where promo optimization really goes.
Three are three major reasons for the slow adoption of promotional optimization in retail.
1. The Downturn. The consumer is reacting very differently to promotional pricing. So there is little to be gained from looking at previous year’s history. This leaves the opportunity to merchants with stable, non-seasonal product offerings.
2. Limited measurements of market basket, stock outs and lost sales. If you are going to measure the success or failure of a promotion, you have to take into account the promotional product pull and the extent to which it was stocked out during the promotion. Most retailers don’t do this well.
3. Lack of a clear buyer. Optimization projects require the intense collaboration of IT and a senior sponsor from the business. IT has been smashed by recession cost cutting. The Business is running for cover.
Promotional optimization will take hold as the economy recovers and seasonal-fashion merchants will have some data that they can work with.
As the march of product proliferation races fast-forward its pace is second to that of the mercury-quick wave of tools to measure the retailer’s promotional progress. The retailer has become a multiple target and he is not too deft in catching and utilizing all the pitches being thrown at him today. Thus his usage, or lack thereof, draws topical and critical questions, but life still goes on.
Why aren’t promotion optimization tools being utilized today? One word: resources. Specifically, the retailer’s IT and Marketing resources.
Existing systems have a very resource-intensive life-cycle: high up-front licensing costs, major investments in technology infrastructure, lengthy and burdensome systems integration, intensive training of a select few highly technical users, and then heavy reliance on those few users to manage an enormously complex process. In order to see mainstream adoption, these systems will need to be less expensive, less disruptive to install, and less reliant on a few highly-trained technical wizards.
I see two possible ways to ease the resource bottleneck. The first would be to simplify the process and the software so that it doesn’t require as much effort to manage. Of course, that inherently might reduce the value of the promotion optimization if it becomes less responsive and flexible. The other approach would be to push some of the work out to more users. This, in essence, is what Kroger has done with its dunnhumby partnership–it has bought access to a couple hundred additional bodies (albeit highly capable, technically proficient ones) to generate, manage, and optimize more promotional activity.
That’s great if you are Kroger and have the huge scale to absorb those costs. But what if you aren’t? Most retailers simply don’t have the resources to generate that level of nuanced promotional activity within their own four walls. Ultimately, the promotion flow has got to come from the brands. And it can’t just be promotion dollars–brands will need to do some of their own lifting to create, target, and manage promotions. Our company has built the tools to make that type of brand involvement easy.
Promotional solutions are tough since each retailer’s needs and objectives are different and within a retailer, their stores are different depending on location and competition.
I worked for years with promotional optimization solutions from the manufacturers’ side and found them to be very helpful for understanding what promotions worked and if they did not work what might have caused the issue. Was it a retailer execution issue or a supply/out of stock issue or was it just a poor promotion? It also took into account the key objectives of the manufacturer/retailer which was key since you wanted to try and run programs that matched everyone needs. From the retailer perspective it can be more challenging since they have so many ideas being thrown at them from manufacturers and outside vendors. I was in a retailer’s office last week and the executive talked about having far too many promotional programs and they needed to sharpen their focus and do fewer, better.
Here are a few suggestions on how to sharpen your promotional programs:
1) Be sure any promotion you do is focused on your core objectives. Just because a promotional program is great does not mean it is great for you. That goes for retailers and manufacturers.
2) Retailers, be sure and share your corporate objectives early on with your suppliers. The better they understand your objectives the easier it will be for them to suggest programs that work for them and meet your objectives.
3) Manufacturers, when the retailer shares their core objectives listen and go back to your team with the information so you can present solid promotional programs that fit.
4) If your core objective is to grow, for example, Private Label, then focus on promotional programs that can help grow private label and share this with your manufacturing partners so they can suggest ways to participate.
5) Break down internal walls that can sometimes slow down programs that will be great for the overall business. Implementing a promotional program that focuses on Private Label may disrupt another internal groups plan and how they are compensated. The way to address this is get executive management involved so all groups are rewarded for driving the business forward using their core objectives.
6) With any Promotional analytics solution, be sure that it allows you to include your core objectives. As stated earlier a promotion may be great (mathematically), but it may not fit with what you are looking to accomplish. Again, that goes for both retailers and manufacturers’.
Feel free to contact me if you want resources to call or have additional questions.
A lot of good comments above. Here’s my view:
A lack of retailer resources applied to understanding promotional effectiveness is certainly an issue. However, I have conducted literally several hundred promotional analyses, including use of customer-level (shopper card) data…and the $ savings potential in making better decisions SO FAR outweighs whatever resources a retailer might throw at the issue, it’s ridiculous. Kroger knows this…it’s how they can pay enormous sums to their ‘outside consultants’.
Other factors mentioned above, such as a lack of software, unclear objectives, data access issues, data quality issues, changing sensitivities during changing economic times, and so on, are all valid inhibitors. But if retailers truly understood the $ involved in promotional effectiveness, none of these inhibitors would really matter.
The TRUE inhibitor is organizationally- and culturally-related, specifically with respect to incentives. For example, I personally have brought Category Managers through results of analyses which clearly show a particular promotion is a financial train-wreck. But they don’t care if they are judged on the VOLUME a promotion provides…even if their bonus includes some kind of profit component. At the end of the day, the institution rewards and celebrates them on the volume a promotion produces. ‘Marketing’ is staff and has no pull. ‘Merchandising’ is all about volume. ‘Operations’ couldn’t care less.
Until the retail organization wholly aligns incentives around the profitability of promotional activities (both short- and long-term), there will not be meaningful progress in understanding promotional effectiveness within the retail community.
I just don’t buy this, unless you are only talking about a small subset of retailers. My company, APT, has more than 40 retail clients and the vast majority of them are using our Test & Learn software to optimize promotions.
I am very confused by the promotion steps described here. I really believe that the “art” should be up front. Most of the facts are given to the retailer by the terms of manufacturer discounts. Performance requirements will often dictate the magnitude of the discounts, the manufacturer’s way of getting even for slotting fees (just kidding). It is the Art of good merchandiser who knows the overall strategy of their banner and the role they want to play in market place that allows the retailer to select from the various manufacturer offers.
The truth is that there are too many moving parts in this whole process to successfully document all the attributes for thousands of items on a weekly basis. When you add custom item lists and different promotion zones, the number of attributes becomes even larger. Histories are fragmented and the impact of exogenous factors such as weather, construction, or a competitor action often make the history misleading. Item numbers and sizes change consistently so that comparing historical data across seasons is often problematic.
I am a “quant.” I believe numbers can help drive better decisions. But just as a set of Calloway’s in my hands will do nothing for my golf score, I don’t believe quantitative software without all the data will give satisfactory results. This is where the art plays a key role and will always be necessary to supplement the numbers. Pricing applications require less input and the history is fairly consistent. There is enough price data so that you can throw out a few oddball results without affecting the final analysis. I don’t think the same kind of analysis will exist for promotions for awhile.
This does not mean automation cannot help with managing the whole promotion process. Capturing manufacturer offers, documenting them to avoid disputes, automating the deduction process for bill backs, presenting the various offers for selection by category, arranging them by weekly ad issue, getting the initial distribution quantities in place, handling the replenishment orders, prioritizing in-store shelf replenishment, etc, can all be supported by an integrated promotion management application. It just can’t tell you what to do.
Most retailers that apply automation to promotions today allow vendor dollars to drive their decisions. I see Promotion Optimization as a more agnostic and goal oriented approach to promotions.
Promotional optimization leverages advanced analytics to allow retailers to identify the very best items to promote at the ideal price point knowing in advance the impact at the item, category, and store level.
Promotion Optimization should be goal driven. For example, if a retailer is trying to drive traffic to the stores, that is usually a completely different set of items than if a retailer was trying to drive profit or maximize vendor dollars. Promotion optimization can consider all of those variables and help the retailer arrive at the best decision.
A good promotion optimization system should also consider the impacts of cannibalization, affinity and pantry loading.
Great comments above.
Apart from the data availability, limited customer analytics, economic downturn, other factors that hinder adoption of promotion optimization –
1.Downstream Complexity – A promotional optimization exercise would typically yield the ideal promotional price and promotional type (BOGO, Percentage off & others), for a customer segment. The challenge is to execute this in stores. Typically retailer’s IT systems grown organically over the years cannot handle this downstream complexity of setting up the promotion at the granular level in the store systems. Typically challenges faced by retailers to execute a targeted promotion ( a promotion for a specific customer segment) are:
a. How do I recognize the customer when she comes to my store – Most loyalty cards issued by retailers are used as discount cards, (one can avail a promotion only when they have a loyalty card). Hence for the cashier at the till, they are segmented as “one who has a loyalty card” and “the ones who don’t.” It’s a challenge to make the information of the various customer segments available to the cashier when the customer is billing. Unless the customer can be recognized, that she belongs to an “X” segment, it’s difficult to execute targeted promotions. Making this data available is a huge investment.
Retailers are looking at mobile coupons to tackle this issue.
b. How do I offer her the right promotion – Once the customer is recognized , how can the right promotion be offered to her and how will the point of sales system be able to handle this differential pricing , is a challenge with hardware and software constraints on most POS systems.
c. How do I make her to buy more – Most loyalty cards are swiped at the end of the shopping trip? Hence the “feel good” factor which triggers impulse purchases is lost, as she has already finished shopping.
2. Different promotional zones and Ad zones: Typically retailers based on their understanding and needs group stores for promotions which are then typically mapped to the AD zones, (clusters which are created based on circulation of a particular newspaper/weekly and others). Creating targeted promotions, would require logistical remodeling to deliver these targeted promotional fliers to the right segment.
3. The “why”: Most optimization tools yield reasonable outputs, however, most category managers/planners are not comfortable with a machine telling them what to do. They need a “why” to the suggestion that the tool make. The tool suggests, coke 1ltr needs to be promoted, but rarely say why – Is it because it will increase sales by x%, or will help clear stock. Most optimization tools lack “the why” of the suggestion.
4. Testing and measurement: Retailers need to understand that the promotional optimization tools do not get it right in one shot. Continuous test and measurement is required to fine tune it and retailers need to invest in building a robust test and measurement process.