3
• High volume, 200 bits per month
In Figure 3, the standard deviation is compared to the
bit size. It shows a strong correlation with larger bits having
much smaller variability. Bits 64 mm [2 ½”] in diameter
and larger are used for longhole drilling alternatively ream-
ing in development drilling. For the longhole bits, two fac-
tors may contribute to relatively stable consumption:
• Drill meters in ore production is likely somewhat flat
driven by the tonne capacity of the downstream pro-
cessing operations.
• Compared to the amount of drill meters, there are
relatively little collaring for longhole bits.
The 51 mm [2”] and smaller bits are used for develop-
ment drilling. Statistically, there is no difference between
bolting bits and face drilling bits. For inventory planning
purposes, the data indicates that development bits should
be assumed to be more variable than the larger bits.
TOTAL CONSUMPTION VARIABILITY
All data by SKU and site were normalized to make the
monthly average consumption 100%. Actual consump-
tion in each month was then calculated as a percentage. For
instance, if a particular average is 123 bits per month but
the consumption in June was 86 bits, that generates a 70%
data point. If the consumption in July was 148 bits, that
data point is 120%. Combining all SKU, sites, and months
leads to the histogram shown in Figure 4.
Monthly variability has a staggering standard deviation
of 42%. Bear in mind that this data has been scrubbed,
only the “high” runners with regular use are kept, and the
mines all claim to have stable production. With Vendor
Managed Inventory (VMI), this data has an inherent bias
towards less variability it can be expected that many mines
have even greater variability.
An appealing explanation for the high standard devia-
tion would be that one month is too short a time period. The
arguments would be that situations with exceptional rock
conditions don’t persist or rigs are repaired and returned
to service. Over- or underconsumption of drill bits in one
month should be compensated with the opposite in the fol-
lowing month. There is obviously some truth to these argu-
ments, but as shown in Figure 5, variability remains high at
a 29% standard deviation even when using a three-month
sliding average.
Based on these variabilities, it would be tempting to
write off forecasting based on historic averaging as a fruit-
less exercise. That cannot be the case – knowing that a bit
averages 50 vs 500 per month has value – but it means that
Figure 3. Standard deviation shown against bit size
diameter, including smaller development bits and larger
longhole or reaming bits
Figure 4. Histogram of monthly bit consumption around
each bit’s average
Figure 5. Bit consumption based on three-month sliding
averages
• High volume, 200 bits per month
In Figure 3, the standard deviation is compared to the
bit size. It shows a strong correlation with larger bits having
much smaller variability. Bits 64 mm [2 ½”] in diameter
and larger are used for longhole drilling alternatively ream-
ing in development drilling. For the longhole bits, two fac-
tors may contribute to relatively stable consumption:
• Drill meters in ore production is likely somewhat flat
driven by the tonne capacity of the downstream pro-
cessing operations.
• Compared to the amount of drill meters, there are
relatively little collaring for longhole bits.
The 51 mm [2”] and smaller bits are used for develop-
ment drilling. Statistically, there is no difference between
bolting bits and face drilling bits. For inventory planning
purposes, the data indicates that development bits should
be assumed to be more variable than the larger bits.
TOTAL CONSUMPTION VARIABILITY
All data by SKU and site were normalized to make the
monthly average consumption 100%. Actual consump-
tion in each month was then calculated as a percentage. For
instance, if a particular average is 123 bits per month but
the consumption in June was 86 bits, that generates a 70%
data point. If the consumption in July was 148 bits, that
data point is 120%. Combining all SKU, sites, and months
leads to the histogram shown in Figure 4.
Monthly variability has a staggering standard deviation
of 42%. Bear in mind that this data has been scrubbed,
only the “high” runners with regular use are kept, and the
mines all claim to have stable production. With Vendor
Managed Inventory (VMI), this data has an inherent bias
towards less variability it can be expected that many mines
have even greater variability.
An appealing explanation for the high standard devia-
tion would be that one month is too short a time period. The
arguments would be that situations with exceptional rock
conditions don’t persist or rigs are repaired and returned
to service. Over- or underconsumption of drill bits in one
month should be compensated with the opposite in the fol-
lowing month. There is obviously some truth to these argu-
ments, but as shown in Figure 5, variability remains high at
a 29% standard deviation even when using a three-month
sliding average.
Based on these variabilities, it would be tempting to
write off forecasting based on historic averaging as a fruit-
less exercise. That cannot be the case – knowing that a bit
averages 50 vs 500 per month has value – but it means that
Figure 3. Standard deviation shown against bit size
diameter, including smaller development bits and larger
longhole or reaming bits
Figure 4. Histogram of monthly bit consumption around
each bit’s average
Figure 5. Bit consumption based on three-month sliding
averages