The following entry is a record in the “Catalogue of Catastrophe” – a list of failed or troubled projects from around the world.
Avon Products – Canada
Project type : Product sales and ordering system
Project name : The “Promise” project
Date : Dec 2013 Cost : $100 to $125M
Ding-dong – Avon calling. According to reports, a significant number of the Avon sales team in Canada will no longer be pressing any doorbells. One of the world’s most established direct sales companies, Avon leverages the multi-level marketing strategy to distribute its products through a network of agents who sell direct to family, friends and their personal contacts.
In such a business your sales agents are a critical link in the chain. If they loose faith in you, then your distribution channel has gone. Unfortunately reports indicate that Avon did indeed loose the confidence of its Canadian sales reps by introducing a new sales order management system. Introduced as part of a global “Service Model Transformation” (SMT) project, the new system was intended to streamline the ordering process thereby allowing Avon to reduce costs and better meet customer needs and expectations. Utilizing a new back-end ERP system and a new tablet enabled e-commerce front end, the SMT system was intended to allow sales agents to use their tablets to showcase the products and then immediately check inventory and secure orders online. Dubbed the “Promise” project the SMT initiative was intended to save approximately $40M per year.
According to a Securities and Exchange Commission (SEC) filing made by Avon on 9 Dec 2013 the project has now been abandoned with a write-off of between $100 and $125M. According to the filing, the system was piloted in Canada but “caused significant business disruption in that market”. The SEC filing reveals that the system “did not show a clear return on investment” and that the decision to stop the project “was made in light of the potential risk of further disruption”.
Reports available in the press indicate that the issues were with the front-end e-commerce components rather than the back-end systems. Basic functions such as logging in, saving orders and checking inventory were not working properly. Furthermore, complaints about usability have surfaced. Users of tablet devices have become accustomed to easy to use apps that can easily be navigated with the touch of the finger. Reports indicate that the Avon user interface was poorly structured and did not meet the expectations users have of a modern tablet app. While back-end systems are typically more clunky and more difficult to use, apps are becoming ever simpler and easier. In yesterday’s applications it was typing, pick lists, fill in forms and multiple stages. In today’s apps, its get the job done in as few steps as possible using graphics, images and icons to make the journey from start to end as simple as it can possibly be. Although the intent was there it appears that design issues resulted in a system violated the agent’s expectations of how an app should function.
The net effect was drastic. One senior sales executive reported that she lost more than 300 of her sales agents as a result of the project (1/3 of her total sales team). Although many of those many have been the smaller volume agents whose lower levels of income may not have warranted battling through the deployment glitches, the SEC filing reveals that the impact was clearly enough for Avon to decide to back-off further deployments.
Note: Although the “Promise” project did not live up to its promise, Avon did at least reduce risk by testing the product in a smaller market before going global. That decision may have saved the business. If you contrast this staged release approach with J.C.Penny’s disastrous all-in approach to its business transformation project in 2012, it is a good illustration of how different approaches can lead to very different outcomes.
Contributing factors as reported in the press:
Quality flaws. Lack of stakeholder analysis and the failure to understand the sale’s agents expectations of an app.
Reference links :