What not to do when you are asked to solve a problem - Part 2

You are given a task by your lead to complete. It's a problem that you must solve. How do you go about it.

A few weeks back, I had talked about how frustrating things go when the solution you provide and what is expected doesn't match. Click here for the article.

From the collective experience of my peers, my team mate and me, I would rather follow the seemingly long steps mentioned as under

1. Go through the task yourself once before you discuss it with your lead (assuming you have received a mail from him) [30min to 45min] 
1. Make a note of what you understand and the questions that you want to ask 
2. Get the following (you may be completely wrong - and that’s the point!): 
• You must understand why you need to do something, i.e., how is your task going to help 
• Insist on this if your lead doesn’t want to answer this the first time 
3. Attempt to first create a rough sketch of the solution expected 
1. Eg if a report (or a presentation) is required, make a rough sketch with the following: 
• The flow of the story --> flow of the slides in a deck 
• A table with various columns that will be displayed as an output 
• The various input options that the report is expected to have 
• Any charts/ graphs with the axes labeled clearly and the type of chart 
2. If an analysis is required, keep a rough note of the following: 
• First and foremost - what do you want to build --> a descriptive analysis or a predictive model or is it a prescriptive framework 
• Other details that must be thought out at the very start (Out of scope for this topic) 
2. Go with your notes to discuss with your lead [30min to 45min] 
1. Make the necessary changes to your notes and confirm your final notes with your lead 
2. Give your inputs now (Get creative - but revisit the Objective if needed): 
• Is there anything else that could be a value add to the customer 
• Can we do this differently 
3. Ask your lead, if you need to confirm this with your client 
• Sometimes, it's better to take the customer's input right at the start - to make sure you are at it 
• Make changes to your notes accordingly 
3. Discuss with your lead regarding the DATA [15min-30min] 
1. If you have the relevant DATA - tables, columns, etc 
• If not, how to get the data (Contact any other team mate? Request approvals?) 
• Discuss definitions and formula to calculate certain new 'measures' 
2. What should be the scope of the data - time interval & any other 'filters' on the data that you must apply 
3. Any special business rule to follow - to change the part of the data 
• You must be thinking that you have taken up two hours already and you haven't started with the deliverable. But trust me, if you have followed this process, your job is half done already. And the rest of the path is as smooth as it can get. And if you haven't followed this process, you are playing a 'finding a black cat in a dark room game'! You might end up with a perfect solution, but not always.
4. Make a note of the 'time to delivery' & understand if it is an ad hoc or a continuous project [15min] 
1. If it is not an ad hoc report, no need to concentrate on automation, etc. 
2. The priority should be to get the work done & validate it for quality 
3. If it is a project, your lead will surely come up with a project plan (including timelines) 
1. Review the same and raise concerns if necessary 
2. Make sure that internally you have a 'milestone' tracker to refer to 
5. Start executing the plan
1. Revisit the task objective and your notes whenever required 
6. Make it a point to review the solution with your lead at the half day mark once 
1. To give yourself a chance just in case things might look even uglier later 
2. Less rework later means faster delivery 
7. Solution delivery to client 
1. If you have already taken the customer feedback of the plan, your job is much easy here 
2. Customer gets exactly what he has agreed upon 
1. Be open to make those 'finer adjustments' to the solution 
2. After all, customer is God ! 
3. You get your peace back! 

Assumption: Your lead has the last word.
However, if you have experienced that your manager [or any other senior member] gives you end of day feedback on the solution and expects the solution to be changed, make sure he is invited to the review meetings with your lead. If this isn't the case, bring this up during your meetings with your manager.
So to answer the question, 'what NOT to do when you are asked to solve a problem?':

Do NOT start solving the problem immediately!

Give it a thought - if you are the kind of person who hate unnecessary rework :)

Let me know your thoughts and share with your experiences about how not 'giving a bit of thought' before you started solving a problem turned out to be a mistake you could have avoided.

Happy Solving
The Analyst Charioteer