Operations Improvement in Small Manufacturing - Five Overlooked Areas to Mine During Tough Times

Even in tough times when they need to maximizeEverybody's got software to sell nowadays. It's
the impact of every nickel, I have seenmanytempting to avoid the complications anduncertainties
manufacturers burning money.of changing or upgrading. We have witnessed a
How?deluge of specializedofferings ranging from erp, mes,
Well, their attention is usually drawn to the obvious,and industry specific CRM systems to process
and companies in tough timesusually tackle those:controland data acquisition systems. And now
cutting personnel, eliminating extra programs, etc.enterprise manufacturing intelligence (emi) to tiethe
However theymay lose money in the long run if theyloose ends together. It can all induce severe
don't attend to other fundamental changes."acronymphobia", not to mentionconfusion and
Here are a few ways among others that I'm familiarparalysis.
with that are less obvious, yet canmake seriousOn the other hand, tough times require that we seize
impact on the bottom line and should be consideredevery advantage possible. How canwe do that when
especially in thiseconomy:I/T budgets are shrinking?
- measuring the wrong quality characteristicsOne approach to consider is allowing an unbiased
- improper sequencing and execution,computerized expert system help withthe decision.
- over investing in the current "as is" environmentThese expert system logic tools can gauge the gap
prior to changesbetween past, present,and future needs and the
- living with business software that is sub optimallycapabilities of the applications in place. Best of all time
implemented or outdated.andassistance can be rented on a per project basis,
- overlooking opportunities to provide closer linkagekeeping the investment modest.
between R&D and productionAt one time, I undertook an application software
I write and share this report so perhaps you, theassessment for a $30M metalsprocessor. The firm
reader, might find a nugget or two foryour ownserved the automotive and allied industries and
business. It's not copyrighted - you can duplicate it toorganized its stockedand in-process materials by
your heart's content andpass it to whomever youattributes (size, weight, length, width, grade, etc.)
think would benefit from it.rather thanby sku numbers. The OEM firms using
SIMPLE EXAMPLE FOR ILLUSTRATION PURPOSESsuch an "outside processor" are essentially
A major candy and confectionery company wasbothsuppliers and customers at the same
computerizing its supply chain to makethings moretime--material arrives, gets processed, and isreturned
efficient. All the information of the raw ingredients,to them.
time to ship, qualitytesting time, etc., were beingEach of these circumstances limited the metal
integrated into one master system. In the processprocessor's ability to conform to morewidely
theymissed one piece of information forapplicable manufacturing software packages. Use of
synchronizing events: seasonal supply chain events.the expert system evaluationtool combined with
They almost missed Halloween! Eventually withsome extra research revealed that an industry
frenzied damage control, they were ableto gear upspecific niche ERPproduct for "attribute based" metals
the production to meet the demand.processors would work better than an initiallycheaper
So, that's an obvious example, and it's such a biggeneric ERP system that could be costly to
SNAFU that it's hardly likely yourcompany wouldcustomize, adapt, and maintain.
make that mistake: your company is not likely toAfter selection and implementation, the estimated
miss it "Christmasseason." But it can miss manysavings for this small enterprise waswell in excess of
smaller practices that can cause bigger glitches. "A$150,000 annually when compared to previous
Stitchin Time Saves Nine," as they say. Trouble atmethods of tracking,organizing, and executing
the headwaters of production can getamplifiedmaterial processing that combined manual and older,
downstream.lessadaptable software. Savings on customization
Here are some illustrations from my experience thatcosts, both planned and unplanned,obtained by
could contain valuable lessons foranyone involved inselecting the specialized rather than a generic
process manufacturing:package added another onetime savings estimated at
1. What you measure is as vital as how well you$100,000. Of course, no one should blame the client
measure it: or... "Productshipped by weight, notfordeclining to run an experimental generic package
volume; contents may settle during shipping."implementation in parallel to verifythese savings!
That consumer warning on packages reminds us ofIs there a suspicion that the capability gap between
every day occurrences: Focusing onone parameterbusiness software and requirementsmay be growing?
while measuring another. Why would you measureNow might be the right lull in the business cycle to
volume if you'reshipping weight? Well, it can getfind out and takeaction.
confusing.4. Needlessly Prolonging Continuous Improvement in
Think of a school bus. What color is it? Yellow, right?Two Crucial Ways
Not exactly. It's actually adistinct mix of trace colorsThe batch flow or semi-continuous process operation
like red, blue, etc, which are added to the yellow inat first glance seems less amenableto finding and
just theright proportions that makes the bus soimplementing improvement opportunities than a
recognizable in dim light or stormy conditions.textbook discrete "widgetmaking" or piece part
The recognition and safety factors, so key to quality,operation you may have seen. Pipes, vessels, steam
stem from less than obviouscharacteristics.jackets, orreactor chambers obscure the easily
Other examples abound. Let's look at anotheridentifiable avoidable waste and "non value
involving formulated industrial products.add"activity, and crews may tend to several activities
At one time I worked with a small maker of low-tackspanning several lines rather thandistinct work areas
hot-melt adhesive that was used tomake tackyor cells. Yet, reasonable bite-sized, digestible
polishing cloths for the furniture industries. One dayimprovements in theseenvironments are feasible.
the company foundproduct was being returned forWhat may make more sense, at the risk of
inconsistent tack levels and poor color andcommitting heresyagainst lean thinking or other
appearance.improvement philosophies, is to move directly to a
The firm needed an objective, targeted measure ofbetter
the quality and effectiveness of thecloth to satisfy"future state", rather than optimizing or dressing up
its large OEMs and its ISO registration requirements.the "as is" state at critical points inthe facility. (I will
The "stickiness" characteristic of the finished productassume a little familiarity with process improvement
was conferred by the polymerimpregnated in thevocabulary here,but the ideas are straightforward.)
surface. Up to that point the company had reliedTo illustrate, it may make more sense to skip the
upon a "thumb feel"subjective judgment of functionwork area neatening and straighteningand "quick hit"
along with a set of deteriorating retained productsefforts called for by work place clean up ("5S" in
asvisual standards for color and general appearance.Japanese buzzwordspeak) and "kaizen" incremental
Employees thought viscosity of the molten blend atimprovements and proceed right to the new "to
various stages was the key.be"process setup. The existing human work space is
Measuring that would call for some expensive, highexternal to the process, after all, andmay even
maintenance instrumentation usedcarefully at preciseprove irrelevant after changes are made
intervals. But I suggested examining the processA hand-blended, barrel and drum scale stamping
more closely with adisciplined experiment designed tolubricant blender that I assisted madethe leap to an
see which process variables are clearly relatedautomated mass flow meter based blending system.
toimportant end product characteristics.This mini-scaleoperation found that neatness and
As it turned out, viscosity was a weak predictor ofreduced clutter naturally followed in the wake of
quality and performance. In contrast,a low tech,thisupgrade once the frantic pace of manual rework
inexpensive surface friction test proved to be farand adjustments abated. It can be moreheartening to
more effective. Thecompany saved $6500 perthe team to make manageable, meaningful changes
device in test equipment investment along with dailyof this sort rather thantrivial or elementary ones or
operatingand maintenance costs associated with theto make pointless changes destined to be quickly
more elaborate testing method.displacedby others.
The subsequent QC testing program that was set upA food microbiological sample testing facility client had
incorporated this and otherstraightforwarda congested arriving samplearea, an immovable, fixed
measurements. These methods kept costs andwalk-in incubator, and other sample prep, scaling,
quality in line and offereddeeper insight into theandsequential repetitive processes surrounding it. The
impact of process conditions and ingredientfacility needed a way to relieve thereceiving area
proportions. Inaddition, the friction test provided achoke point and get samples processed faster to
vital element of the objective "processservice critical customersbetter. Through our help
control"requirement of QS/9000 (now 2000)they found that adapting continuous production flow
guidelines.techniquesto "pull" products downstream, eliminate
This new approach enabled the firm to obtainunnecessary unproductive movement, andproperly
compliance and satisfy vendor certificationand suppliermanage incubator capacity and timing provided
audits from large key customers and to qualify foralmost instant relief to thereceiving area. Business
new business.was solidified and billable volume increased, while
Is your firm, large or small, simple or complex,workers gothome earlier and avoided burn-out. Time
measuring the right things for the rightreasons?spent on sprucing up the receiving or first
2. Haphazard Sequencing of Orders and Processesstageprocessing area beyond removal of safety
Causes Waste, idle time, delayshazards and some basic mistake-proofingwould not
Everyone knows the aggravation of living with thehave yielded comparable benefits and would have
limitations and constraints ofprocessing equipment.postponed genuineimprovements.
Only so much quantity and variety can be producedSimilar benefits were obtained in a small continuous
in a givenperiod. What is less obvious is that there ispasteurized fluid beverage facility.
often still room to adjust the sequencingHere, ironically, manual clean up paraphernalia created
andconditions to maximize throughput. Thesemore clutter until better plannedcontinuous runs of
adjustments can even go a long way towardfulfillingproduct reduced the frequency of changeover and
the "lean" goals of less waste and idle time. Thisdisruptive cleanupsteps.
requires at least a littledisciplined logic and spreadA second cause of prolonged improvement initiatives
sheet figuring over and above what a visualis attempting to painstakinglystreamline every
scheduling toteboard, grid diagram, or other types ofidentifiable sub process node in a larger production or
crib sheets can provide.paperworkprocess regardless of its impact on the
In one fluid agrichemical operation I observed, thewhole. More often than not there are crucial
purchasing manager, the shiftforeman, and a plannersubprocesses that govern the performance of the
scheduler all had their own schedules. No scheduleentire process--either in rate of production,build-up of
wascompared with any other until after manywork-in-process queues, or control of quality. Zeroing
production orders were already launched.in on them will helplimit the risk of creating locally
Which one prevailed? You guessed it, a fourth in-plantoptimized processes that don't contribute to
"on the fly" compromiseschedule based on hunches,overallperformance. In the lab case above, speeding
incomplete information and last minute corrections.up the rate of completion of final testingsteps that
Other examples can be found:occurred after the incubator stage, while possible,
Consider two consecutive interdependent processingwould have only tightened afew isolated downstream
stages with separate processdurations that aresets of procedures. Improving the batch handling of
somewhat predictable in advance. The sequencedsamplesthrough the incubator, on the other hand,
stages might involveagitating, heating, curing, drying,made a significant impact on the overalleffectiveness
and the like. The duration of each will varyand efficiency of the lab.
dependingon the raw material in question, theEven more dramatic examples occur among plant
conditions chosen, even the weather (as infloor operations where certain stepsinvolve
conveyingand melting wax or other substances). Thisbottlenecks or "rate limiting" steps. Finding all
means that for a given job one process maytakeimprovement opportunities is agood long term
longer than another, while the reverse may be truecontinuous improvement objective, but only a critical
with another job.few may yieldsuccess in the short run. Are there a
A producer of dispensing machine beverage powders"significant few" areas for improvement
was experiencing order stack upsand random waitingbeingoverlooked while the "trivial many" are under
and idle periods as the sequence of differentstudy? It may pay to take a look.
materials wereprocessed. As part of the general5. Missing Ways of Bridging R & D and
process overview, I learned that the company had aProduction Activity and Data Domains
twostageblending and spray drying process. In thisWhile discrete manufacturing environments seem like
process some raw materials dissolvedbetter thannaturals for integrating design andmanufacturing what
others, while some finished blends required slower,with CAD, shop assist drawings, and tightly
gentler drying cycles.orchestratedengineering change orders, the situation
In such cases a good way to sequence the jobs isis murkier in process manufacturing.
to find the job with the shortestpredicted processComplexformulations and lab scale chemistry, biology,
time for either stage. The job is then queued up firstand material science can lead to manyloose ends on
if the first stage isexpected to be shorter than thethe way from R&D to production such as
second. If the second stage is to be the faster one,separate bills of materials,recipes, or approved
the jobis queued up last.formulae, not to mention separate quality and
The process is repeated with the list of jobsperformancemeasures. Scale-up pilots that show
remaining, filling the sequence in from bothendsunexpected cause and effect relationships
toward the middle until all jobs are sequenced. Anamongvariables and observed quality or performance
optimal sequence minimizingtotal process time willcontribute more complexity.
result.A slight modification of this procedure usingOne confectioner I worked with experimented with
correction factors can be used in cases where themultiple methods of monitoringchocolate liquor
second stage can start before the first is complete,viscosity and solids concentration for the purpose of
as with split orders, drawoffs,and partial batchcontrolling andadjusting end product characteristics. I
transfers.worked with them to construct an open loop
This approach does not require expensive ordatacommunication link between the lab and the
sophisticated software or more thanliquor and sugar tank farm areas to
visualcomparisons and spread sheet level calculations.makeadjustments in full-scale special product runs.
This approach also proved useful in environmentsThis approach avoided the typical delaysand lags that
such as a melt and mold sequence for wax slabnormally require chasing and catching already
preparation where melting and molding steps canprocessed product to makechanges. Eventually, the
alternate as bottlenecks,de pending on the grade ofclose link-up allowed R&D to fine-tune the
wax and other conditions.in-process recipe toobtain a richer, smoother mouth
Note that the objective is not better equipmentfeel product that commanded a better price and
utilization for its own sake, but betterthroughput ofprofitmargin. Later modifications were made to QC
job orders per time period--what generates revenuetesting to make spot checks on the sameproduct.
for the business aftersunk costs have beenThat effort brought QC and R&D into closer
committed. (No one has ever been paid directly foralignment on their methods andpriorities, bridging a
howuninterrupted their equipment usage is, unless it'sgulf almost as serious as that between R&D and
to settle a bet!)production.
Are there process sequences in a facility you knowCareful combinations of procedures and technologies
that could be more productive? Theremay be a waycan bridge functional gaps andcultural "brick walls" that
to get that productivity without equipment upgradescan seem insurmountable or are simply "part of
or expansions.thelandscape". Might there be a few applicable to
3. Ignoring growing gaps between aging applicationyour environment?
software features and needs