This blog presents five key lessons about how to manage an outsourcing program based on some infamous projects that went wrong. I should underline that outsourcing normally goes well, and that there are many good examples, but there is a time and a place. It is certainly a risky strategy for niche, mission critical applications and you should never allow it to dilute your in-house expertise like Boeing did. Before I start to sound “holier than thou,” this topic is personal to me. When I ran product management for a global communications company, I inherited their first outsourced software development, it failed for all the textbook reasons listed below, including lack of definition, product specific knowledge and governance. When the project is mission-critical, you have less time to turn it around.
Eight infamous outsourcing and consulting debacles are listed below, public sector cases dominate the list:
|Project||What Went Wrong|
|Boeing, 737 Max: Fatalities||Following fatal crashes last year and in 2019, the Boeing 737 Max became the new flagship for software disasters, with badly governed outsourcing to Indian firms being highlighted in the findings. While the 2013 issues with grounded 787 Dreamliners were hardware related, the latest incidents are related to failures of mission critical software, namely the anti-stall system. Boeing drove too hard for cost savings through outsourcing, devalued its own expertise and mismanaged its partner relationships.|
|The project aimed to centralize UK-wide patient care records, connecting 30,000 doctors’ surgeries with 300 hospitals and the wider healthcare system. The public would have online access to their records. In 2002 an underqualified leader was appointed. A panel of big firms took on different branches of the £12.7B project. By 2009 the program had failed on most fronts and was castigated for poor security and patient privacy. In 2019 many UK surgeries are still running MS DOS based systems.|
|US, Indiana Welfare System: $1.4B||A $1.4B contract to outsource welfare payment ended in spectacular failure followed by a protracted, public-space lawsuit.|
UK, Universal Credit: £837M
|Outsourced to four global IT giants and intended to simplify the welfare system, Universal Credit arrived after 8 painful years at a cost of £837m and didn’t work. People suffered losses beyond the prevailing impact of governmental austerity measures. It will be replaced by an in-house system.|
|Australia, Queensland Health: AU$1.2B||This started out as a AU$6M payroll system project. A combination of incompetence, political interference, a seemingly proactive lack of purchasing governance and absence of remediation saw it spiral past the billion mark. The contractor was paid in full and avoided any repercussions.|
|US, Texas Data Center Rationalization: $863M||This project aimed to consolidate data center operations for 27 State agencies into two DCs. By the first contract suspension in 2008, only five agencies’ systems were migrated with five in progress, back-ups weren’t being performed. The project deteriorated rapidly to 2010 with the original 2 DC goal not achieved and critical errors in almost every aspect of the solution design. New suppliers were appointed on a $1.1B contract in 2012.|
UK, Border Agency: £1B
|In 2003 “e-borders” aimed to control illegal immigration to the UK after border exit checks were scrapped. With £1B spent and the goal still not achieved the project was finally cancelled in 2014 with 50,000 rejected asylum seekers unaccounted for. The contractor won a £224m suit for unfair termination.|
|UK, BskyB/EDS: £318M||The consultancy was out of pocket here when Rupert Murdoch successfully sued them for damages and expenses. Previously, a £50M CRM system was announced to shareholders before the project inception. A consultancy was rapidly selected for a barely-defined project and a high-risk rapid application development [RAD] approach was adopted with disastrous results. The work was eventually performed in house.|
Five Lessons from Failed Projects
1. Agree deliverables, responsibilities and metrics upfront.
Every example above had problems with the definition of projects, key deliverables and the demarcation of responsibility. In most cases the contracts were flawed and poorly governed, allowing the suppliers to make extraordinary profits for poor or non-delivery. Successful projects describe the project’s end state and define incremental deliverables. Key-stakeholders are appointed with specific responsibilities on all sides and a service delivery framework is established. The experience level agreement [XLA] is a concept that is starting to replace SLAs and KPIs, that can drift out of sync with a project and suggest success even when things are deteriorating. The XLA provides a way to monitor on-going success in dynamic development environments.
2. Commitment is required on all sides
High staff turnover can be a symptom of deteriorating projects. A full change management program should be established from the start to keep employees informed and engaged. Changes of leadership and corporate acquisitions are also events that need to be managed in order to maintain commitment and focus. HP’s acquisition of EDS exacerbated the problems of more than one project listed above. Any handover process between individuals or companies should include a detailed briefing on the importance of a successful completion. In environments where administrations change regularly, efforts should be made to either register commitment from all parties in advance or to scope projects such that they can deliver value within a defined timeframe.
3. Maintain flexibility
External factors such as technology development and regulatory constraints can mean that best practice changes dramatically during a project’s lifetime. Perfect execution is an impossible dream but by monitoring the environment, running risk analyses and making informed decisions, the best approach can be adopted to achieve optimum outcomes. JP Reis advocates for the flexible efficiency of an Agile® approach but while rapid software development is a trend, the BSkyB example above shows how rapid application development can be destructive when objectives are poorly defined and governed. NB. Maintaining flexibility does not mean introducing scope creep, which can be an indicator that goals have become unclear and money is being wasted.
4. Don’t waste your expertise
Boeing were so caught up in cost cutting that work was taken from their expert engineers and handed to $9 per hour IT graduates on the other side of the world. Even when the consultants were stationed in Puget Sound over the road from the client, work would be sent back and forth repeatedly due to a lack of understanding. There is value in outsourcing expertise from strategic partners who are industry focused and able to play to their strengths; the “jack of all trades” consultancy is becoming less popular. Communication is an ever-present factor in achieving success so there is no point erecting language, culture and time zone barriers between client and partner. In the case of Boeing, they could communicate what to do but not the best way to do things, this is part of the danger of outsourcing mission critical work. On-shoring, near-shoring or at least right-shoring are the way to go and ideally, consultants will be in the building, integrated with the project team. These approaches are still possible with overseas consultancies.
5. The price is right
Going too far to save money is likely to backfire. Be wary of replacing expertise with basic competence, or worse. The inefficiency will undermine any upfront savings. Be suspicious of a deal that seems too good to be true because you may wind up with a partner that becomes apathetic, lacks focused innovation or whose main aim is to bring new revenue streams to a project. Likewise, don’t overpay or you may get a complacent supplier who milks their cash cow and becomes less motivated by project outcomes. Run a thorough tender, establish benchmark pricing and carefully draw up contracts that define what is being paid for. Make sure there is satisfactory pricing governance and be careful to ensure that incentives serve the best interests of both parties.
Outsourcing can be the key to accessing specialist resources and it can reduce costs and commitments. Companies like Slack, WhatsApp, Skype, Bank of America, Barclays, Citi and JP Morgan Chase all boast of successful outsourcing programs. It just has to be done with attention to detail and careful risk management, preferably with a strategic partnership approach that suits stakeholders on all sides. As the software world becomes larger and more complex, the professional management of outsourcing partner relationships will only become more important.