Sleepless Nights for Select Comfort – Rough ERP Go-Live Leads to 25% Hit on Stock Price

During the first week of October 2015 Select Comfort Corporation (SCSS) went live with a big bang ERP implementation.  Three weeks following the go-live the CEO announced during the quarterly earnings call that things were going well with the ERP implementation and that the company was returning to normal operations.  Fast forward 3 months to the recent February 11^th quarterly earnings release and a different story emerges.

Same store sales fell 30 percent, customers experienced delivery delays, and the company experienced higher-than-normal return rates and order cancellations.  Company sales were off by $83 million dollars in sales.  Earnings were $21 million lower than expected.  The struggles with the ERP system implementation were identified as contributing to 43 cents a share in lost profitability.  While the worst appears to be over, the company is expecting to take a hit of $40 to $50 million dollars more in lost sales in the 1^st quarter of 2016.  With the earnings release the company stock price fell 25 percent, destroying $250 million dollars in shareholder value.

Select Comfort, the maker of the Sleep Number Bed with just over $1 billion dollars in sales across more than 475 stores across the US, is no stranger to ERP implementation dust-ups.  In 2008 the company announced its sudden decision to stop all work associated with an SAP ERP implementation. The company had come under pressure from the Clinton Group to quit the project citing it as both hugely over budget and behind the schedule.  The observations of the Clinton Group also identified that the company had taken up the pursuit without the assistance of a system integrator.  Stopping the project resulted in the axing of 120 jobs and a write off of over 25 million dollars in capital investments.

While no specifics about the project have been published by the company, our research indicates that the program involved the implementation of Oracle R12 and Siebel products.  The firm spent more than $30 million on the program in 2015 with estimated spending on the project nearing $100 million over the last 3 years.  The big bang implementation in October linked 3 manufacturing plants with 20 distribution hubs and over 475 company stores.  We could not identify a system integrator that was heavily involved in the project.

Under pressure again from an activist investor, the spending on the ERP program was identified as a failing of management.  The company appears to have been boxed into a go-live at the start of the 4^th quarter.  October is typically the only month that companies can go-live in the 4^th quarter if Sarbanes Oxley regulations are in play.  Select Comfort also would likely experience seasonal volume increases in the 1^st quarter in support of President’s day sales.  This could have put the company in the position of October or April for the go-live.  A six month delay would have added more grist to the mill for the activist investor so October likely was selected based upon constraints rather than as a measure of readiness.

So what went wrong?  We have attempted to piece together our assessment of what likely went wrong through an analysis of the company’s earnings transcripts, comments left on Glassdoor, Consumer Affairs and LinkedIn, as well as letters to shareholders from Select Comfort and the activist investor Blue Clay.

• Lack of a quantifiable business case – identified by Blue Clay, it appears that Select Comfort lacked a business case that identified a specific ROI and the internal metrics that would be specifically impacted by the ERP implementation. The ERP implementation business case, if used properly, provides a guide to the project team in making the calls on trade-offs that are required during process designs.  Without a formal or active business case, significant design compromises are often made resulting in overly complex designs.
• Focus on budget and schedule as trade off against operational continuity – in her 3^rd quarter earnings remarks, company CEO Shelly Ibach announced the implementation, placing an emphasis of being on budget and on-schedule. As I mentioned before, the company was under pressure from the activist investor as well as boxed into October.   But what compromises were made to achieve the budget and schedule targets?
• Big bang approach to implementation – high risk / high reward. It is likely based upon the makeup of the legacy systems that a big bang implementation was the only cost effective way to convert.  However, these types of implementations should only be attempted with a significant amount of testing, training, and contingency planning.  Faced with the October go-live date, it is possible that some of these mitigating tactics were sacrificed to achieve the date.
• Failure to validate supply chain integrity – in her 4^th quarter earnings remarks, Ibach comments on the loss of demand signals from the hubs to the manufacturing plants. There are a number of reasons that these demand signals could have been lost including:  master data integrity, failed business process to convert forecast or customer orders to demand, or the inability of management to interpret what was being presented as demand.  Testing supply chain integrity is a non-trivial process, however, given this was a big-bang implementation; a test should have been executed.  UpperEdge has no evidence that these tests were not under-taken, but the evidence would suggest that if they were, they were not well constructed.
• Insufficient monitoring of business process performance. Business process monitoring is an area that is largely missed in go-live planning.  Business process monitoring involves actively reviewing internal business performance metrics vs. expectations, tracking system exceptions to assure that they are being resolved and corrected in a timely manner, and monitoring process compliance to gain confidence that work-arounds have not been implemented that can result in creating a bigger mess to clean up.
• Inadequate contingency preparations. Nobody likes to think of the possibility that the implementation could cause serious problems.  It is equally painful to think about the actions that might be required to recover in the event a major problem did occur.  In the case of Select Comfort, it does not appear that the company had the ability to manually clear the backlog that had been created.  It is also not apparent that the customer service team was adequately informed and trained on how to deal with the troubled implementation.

So clearly there appears to be a lot of things that could have been done to reduce the pain that was brought on by this implementation and, in retrospect, spending a few million more dollars on contingency planning and testing would have likely been worth it.  However, it is also important to point out that there are a lot of good things that Select Comfort has going for it.  To name just a few:

• The CEO clearly has a grasp of the situation and is actively engaged in monitoring the recovery.
• The CIO is a seasoned operator and long-time employee of Select Comfort. It is highly likely that he is establishing the appropriate business priorities for the allocation of project resource.
• The company has identified the areas of process complexity that can be streamlined leading to future operating efficiencies.
• The company has been transparent with shareholders regarding the problem and the actions they are taking to resolve.
• The team is likely not bloated with system integrator resources and capable of resolving the issues without incurring significant costs.
• The big bang approach likely means that the major pain is over.

While the next quarter is likely to be challenging, my long-term feeling regarding Select Comfort is positive.  The marketing team will have its short-term challenges of overcoming the damage the implementation may have caused to the brand, but the company is likely now positioned to reduce lead times, lower operating costs and support margin expansion.