At the LISA forum in Boston, I spoke about the challenges of managing teams in times of company transition. The following article, based on my speech, was published in the industry newsletter.
In early 1997, Corel shut down part of its localisation unit in Dublin (the Business Applications Unit) and relocated it back to Orem, Utah. The consequences of that decision had a dramatic impact on the Dublin office which was to continue its focus on Graphics localisation.
Staff morale was seriously affected and over the course of 12 months, the company saw its staff turnover rise to approximately 40%, losing an average of two people a month. Corel also found it difficult to recruit experienced staff as people felt uncertain about its future.
However, by late 1998, Corel not only halted the decline but had reduced it to less than 1%. Furthermore, its productivity in terms of ability to sim-ship products went up from 1 language in 30 days (Draw 7 in 1996) to 5 languages within 30 days (Draw 9, 1999).
How did we do it?
Essentially, we needed to embark on a significant change program that went beyond a simple tweaking of organizational processes: we needed to put more focus on our organizational development.
In the 1960's, the Shell organization developed a change framework which focussed attention on operating the main levers of change: Values, Owners, Customers, Actors, Transformation and Environment (VOCATE). This is a useful framework for understanding how we went about our change.
Values
We began with a single vision for our people, which we encapsulated in the following mission statement: To develop our people to a professional level, where everybody wanted to hire them but to create an environment where nobody wanted to leave.
Owners
This involves deciding who owns the problem. In this case it was simple enough: as Localisation manager for the Graphics group, I owned the problem and needed to do something about it.
Customers
Again, this was fairly clear: the customers were our staff. Our mission statement came about after we examined the issues raised in exit interviews from people leaving and from talking to our existing staff. There were a number of issues that needed to be addressed by the change program.
The most significant issue, from the staff point of view, was the lack of an obvious career path in localisation. At the time, we only had four key job titles: Project Manager, Team Lead, QA Specialist and Localisation Engineer.
As a manager, I had two concerns. First, that there was a 'dilution' of professional localisation knowledge within the company. This knowledge dilution had come about partly as a result of our recruitment practices (in the early days, greater emphasis was placed on hiring people without direct experience in the business) and the fact that people tended to move quickly within the industry without absorbing a broad range of fundamentals.
My second concern was that I felt there was a lack of ownership throughout the localisation group for 'taking charge' of issues. I had about three or four people who were always pushing initiatives, but I wanted more people to realise their career paths really depended on how they contributed to the organisation.
Actors
In the change framework, actors are considered those who are actually involved in the change. In some cases it will be a few key people - project champions - but in this instance I needed everyone involved, although I would rely on a few key individuals to drive the initiatives.
Transformation
This is the cornerstone of the whole change process. It began by establishing the mission statement of what we wanted to achieve. Next, we held an offsite for all the localisation staff. The theme of this Offsite was "Professional Development within a Professional Process" and it focused on encouraging the teams to concentrate on raising their standards both in personal development and in the way they approached their jobs.
The objective was to raise morale, provide a strategic direction and to show the teams that there was plenty of opportunity within Corel.
It involved key individuals who had already shown the sort of spirit and initiative I wanted to instill in the organization. They gave presentations on the work they had done, placing greater emphasis on the fact they had gone ahead and done this themselves. A perfect example was our QA Automation lead, who had worked on this in his own time and demonstrated it could work in localization (we had tried before, but had failed, thereby writing off QA Automation as too time consuming to be useful).
To reinforce the message, we prepared and handed out Go Further "pocket cards" to each employee as a reminder of our goals and how they could contribute to them (see below). Initially, I thought people might scoff at the cards but they turned out to be one of the most useful tactics in the whole change process. Frankly, I was surprised how many people ended up sticking them up in their cubicles.
On one side of these cards, we posed three questions:
On the other side the card, we articulated four key elements of our strategy, which was to focus on training, QA automation, improved test script and file tracking.
We have also redesigned the localization engineering and QA career paths to provide more promotional and development opportunities for staff.
Primary changes included: the introduction of an Principal position and an Associate position in Quality Assurance; the introduction of an Engineering Specialist (midway between an engineer and Senior Engineer) to encourage the technical development of engineering staff; a QA Process Specialist position which focuses driving process change across localization operations.
Another element of the strategy involved the introduction of a Mentor programme into Quality Assurance. We realized that it takes approximately one month to bring a new employee to the performance levels we require and the Mentor programme gave us a way to accelerate this induction. Essentially, a Mentor is chosen from our current team and assigned on a one-on-one basis to the new hire. Their role is to ensure machines are set up, teach them how to use our tools and processes, and to address any difficulties they are having with the job (such as machine setup, understanding the network mapping and so forth).
Finally, we needed to address the issue of 'knowledge dilution'. We did this by pairing up individuals and assigning them particular localisation topics to research and present to the rest of group. So, for example, we had presentations on code pages, date & time and holiday formats, regional settings.
Environment
In considering change and the environment, you need to look both at the internal environment and the external environment. Although we had a pretty fair idea what issues existed internally, we decided staff had a large part to play in contributing to the development of that environment.
Following on from the staff presentations on localization topics, we gave them the task of presenting recommendations on what they would change in their environment. This raised a number of 'soft issues' such as concern about an unfair distribution of workloads and timekeeping, and a number of 'hard issues' such as improvements in the general process (such as improved bug logging procedures and test script management and tracking). Ultimately, these issues were addressed.
The external environment was another challenge. Publicly, there was uncertainty about Corel's future: the stock price had fallen dramatically and we had shut down a division. So it was no surprise that we were having difficulty recruiting. We also found, from talking to recruitment agencies, that we had a reputation for lots of overtime and that we were slow in processing job applications.
First, we addressed our recruitment practices. We established a policy of "recruit slow, hire fast". In other words, we placed greater emphasis on individuals who showed a capacity for change, rather than just direct experience in the business, which meant we interviewed quite a lot of people before choosing a candidate. However, once we decided on a candidate, the goal was to finalise the contract and make the offer within 24 hours.
We also increased the number of people who interviewed the candidate, designed a new set of interview questions and used the interview as a opportunity to 'sell' Corel.
Selling Corel as a career opportunity was an essential part of the strategy and we embarked on a series of visits to the various technical colleges to present the company, our the products and the job opportunities available.
The final part of the strategy in managing the external environment was to create a source of supply for human resources. Dublin has a lot of localisation companies and there's a lot of movement between them, but very few new entrants. We needed to address that.
We took part in an initiative, championed by Anna Brady (ex-Symantec), which involved working with the local government employment agency (FAS) to design and implement a Localization Quality Assurance Course aimed at training the unemployed. Over the past 18 months, we have sponsored and ultimately employed about 12 people from this or similar courses. (To avoid the 'dilution of knowledge' factor, we use a rough estimate of no more than 10-12% of the staff should come from the 'inexperienced' pool.)
Where we are now
Following on from these initiatives, we have had some dramatic results. Immediately after the 'offsite', two individuals came into the office and admitted they were out interviewing but had now stopped as they realized they were in charge of their own careers and there were opportunities in Corel.
Our turnover dropped dramatically, so that from June 1998, it virtually stopped. At the moment, it is running at between 1-3%.
Our productivity has gone up as well. When we shipped Draw 8 (two languages shipped within 30 days) we had approximately 12 engineers. For Draw 9 (May 1999) we had 9 engineers and will ship 5 languages within 30 days. Also, the average release delta for all languages (10) has been reduced by over 50% in the past three years, from an average of 147 days to 66 days.
And all this was achieved without a corresponding increase in overtime. On average, the amount of overtime per head is approximately 30 hours spread over a period of 8 weeks leading up the final stages of the project. This compares with roughly 100 hours per head in the same time period for previous releases.
However, perhaps the ultimate proof that the strategy worked is to see what happened in late 1998. Corel decided to shut down its Orem facility, moving development to Ottawa and relocating the Business Applications localisation to Dublin! This meant integrating existing staff into a new structure, recruiting new teams, and setting aggressive delivery targets. All of which went without a hiccup. In fact, the first delivery of the new Business Applications team is a sim-ship of the WordPerfect suite in 3 languages, all within 30 days of the English product.
