You forecasted the demand for this month but the numbers were wrong. Where do you start your investigation so that next month's numbers are accurate?

your Question is related to short forecast adjustment.

the 1st two thing i would investigate is :

1) functional Check:

segregation of the Demand & identifying the wrong SKU/product Model which has no "forecast accuracy" and which made my forecast wrong comparing to actual demand. a justification related to this Product should be Clear Like (shortage, not ready to sell, EOL, ....)

2) strategical Check:

i compare YTM (year to Months figures, with consecration to GR (Growth Rate). is My forecast for this Month compared to Last year Month x by GA growth rate is normal & justified? is it way higher or way Lower?

I will check first the Sales forecast for the coming month, compare it to the available stock, taking in consideration both the market situation with the company cashflow budget projection in order to maintain the stable balance of the company financial status.

The rate of demand is sometimes different from what we expect so to avoid this problem we need to check the following data : 

1.  On what was adopted in requesting this quantity . 

2.  The nature of market because it is differ in somewhere .  

3.  The nature of products that I had been purchased and Does them been accepted  by the consumers .  

4. Visibility of items and the volume of competition .

During investigation I would first re-check my calculating parameters, like expected sales shared during request vs actual, along with product stock levels shared to be maintained at warehouses.

Then double check product delivery details from supplier's end.

It is best practice to check the Data for previous 6 months and investigate the root cause of the problem.

Have customer complaints been resolved satisfactorily and regular customers resumed placing orders?  Were seasonality factors properly considered?  What tactical or strategic moves were recently implemented by our competitors?

A change in the numbers that were originally forecasted requires immediate notifications to all groups/departments that maybe impacted. Analyzing the root cause of the mismatch in numbers should be done on multiple levels. It could be inaccurate data from the warehouse. A change in the schedule for the project underway. It may also be suppliers issues.

Purchasing and Logistics are very much intertwined, If the issue is not related to suppliers underperforming, it is usually lack of informatio sharing between logistics and purchasing. 

A good Enterprise software is essential to eliminate all kind of errors including but not limited to human errors.

If a supplier informs me of a delay in shipping, first of all I would work on knowing root cause for the delay and if any solution can be applied to lessen the impact. As a Demand Planner and Procurement Professional, I would immediately inform all the stakeholders for this delay and if possible,

A forecast is a prediction of what is going to happen. however forecast makes organization to understand how it moves forward. in a situation like this, yes i would check whether the numbers are right at first and then i will look at the sources of information collected for the forecast and should need to verify them. a forecast can not be 100% right but with a sound understanding the trend it can match at least 90%. 

consumption at floor level

I would go back to drawing board and see if there is a mistake in calculation . secondly I would analyze if there any change in demand or other market forces.