Lately in some popular science magazine the Big Bang theory was discussed together with speculations on possible ending -scenarios of the universe. One such scenario was called the “Big Whimper”: the universe would slowly come to a standstill.
Reading about Big Bangs I cannot help but think about ERP system implementations that have in their project plan a go-live milestone based on switching overnight from the old to the new system. These are the projects that too often end in a “Big Whimper”.
Big Whimpers happen when the Big Bang scenario has been taken for granted and not been properly reviewed.
“Of course we switch from old to new in one go, otherwise we’ll need to maintain two systems and keep the data synchronized”, “It is too expensive to make temporary interfaces with the old system”, “People will get confused when there are still two systems used in parallel.”
Such arguments you hear when you question the sanity of a Big Bang go-live, and surely these are valid points.
But what happens when “D-day” arrives? Very likely the project is late and we have the first crisis meeting with the steering committee. Everybody promises to commit to the next go-live date, and overtime rates and stress levels go up. Comes D-day number two and another crisis meeting. This time the project manager is better prepared and presents a proposal how to go partly live in department A and roll next month out to the other departments. Everybody happy: pressure of the project, we can declare victory, and start to get some benefits.
Why was it unreasonable to propose at the project start a phased-in go live, which became suddenly feasible under the pressure of creeping deadlines?
The main reason is that no sufficient scenario planning is done regarding the go-live phase. Everybody is very busy to make the project contract, buy licenses, ramp-up the project organization, and start making the blueprint. The go-live is just a milestone somewhere in the distant future. It is felt that any other go-live scenario than the Big Bang will only complicate things and make the project duration too long.
This go-live phase should be thoroughly planned at the very beginning. Do a time travel exercise, review the different possible outcomes, and assess these scenarios with costs-benefits and risk management tools. For example: can we really afford the 50% risk of not invoicing our customers for one month, can we do some more pilots, do we have a rollback plan, …?
Where possible, try to avoid the Big Bang go-live scenario, because they are by nature high risk. It still might be that after careful review the Big bang scenario is more feasible, but at least you are properly prepared and it won’t turn into a Big Whimper!