INTRODUCTION

Flash Inc. is a United States-based semiconductor designer and manufacturer. All its manufacturing operations are based in the United States, but it has customers worldwide, especially in Asia, wheremanycomputerand cellphonemanufacturersoutsourcetheirFlash chipsetsprimary production.Flash'schipsetsare usedinthreemainmarketsegments:personalcomputersandlaptops,Internetof Things(IoT),such as home surveillance cameras and thermostats controllable through applications on a smartphone, electronic toll collection systems, among others, and data centers.

CURRENTPERFORMANCE

InpreparationforFlash's2022annualreport,thechief financialofficer(CFO),RickRosenfeld,hasaskedthe treasurer, Theresa Whitman, and her team to perform acomprehensive varianceanalysts throughselling pricesthe comprehensivevariance analysis that explains theprimarydriversofchangesinplannedrevenueandcostsofproduction.Thisrequestwasmadebecause theCFOandchiefexecutiveofficer(CEO)neededtoexplainFlash's2022performancetoinvestorsandanalyststhroughaconferencecallbeforeitissubsequentlyincorporatedintotheannualreport.

Netincomewaslowerthanplannedfor2022byabout8%or$144million,andsaleswerelowerthan plannedbyabout3%or$429million(seeTable1).Flashsetsitssellingpricesattwotimesthecost ofproduction.Therefore,CostofGoodsSold(COGS)isapproximately50%ofsales.1Changes incostsareveryimportantsince inefficienciesinproductioncostsarereflectedin pricesthat can influence demand.

Theresacallstheassistanttreasurer,JohnBurns,andthenewseniorfinancial analysta recent,AntoineBanks, arecentgraduateofatop-ranked master business administration ofbusiness administration(MBA)program,intoheroffice to assign them the variance analysis task.

Theresa:It's that time of year again, but unfortunately, we have less than a week to complete all ouranalysesandrelatedPowerPointslides.Wetypicallyhaveaboutaweekandahalf.Justhow the calendar turned out this year, what can I say?

John:Luckily, we now have help. Antoine will take the lead on this, and Antoine, I'll be there every step of the way to help, so please don't hesitate to ask if you have any questions.

Antoine:I did plenty of variance analyses in my undergraduate and graduate business school program focus a . Beyond that, I didn't do much of that in public accounting before grad school.

John:Well, it's not very different from what you did in school. The main difference here is we tend tofocus abitmoreontherevenueside, butthecostsare inputs, so Rick likes to be able to explain things in terms of Flash's products and alsoimportant.Wehavemultiple products and inputs, so Rick wants to be able to explain things in terms of Flash's products and market segments. Keep those points in mind as you do your analyses.2Also, management is concernedwithdeviationsfrom the planorthe prioryear.Ourproductcostsarealwaysabout50%ofrevenue aswemainlydoublethecostofproduction.Ascostsdecline, weseamlesslypasssomesavings on to our customers.

RickandTheresaespecially where the wanttounderstandwhere therevenuegrowth orlackthere of iscoming from.Howmuchofitiscomingfromthehigher growthIoTsegment vs.theold-line PCsegment,etc. As for profitability, they want to understand the respective cost drivers for COGS and how Flash'sinputcostsvariedvs,theplan(orbudget).Particularly,inlightofMoore'sLaw,3theywantto understand what's driving the higher cost of the new generation (i.e., eMLC) chips.

This is important because if our costs are too high our prices become too high and that's probably driving some of our sales declines.

Antoine:Gotit!I'llclearmyplateandgetonitrightafterlunchthisafternoon.

Theresa:Antoine, a final hint as you prepare your analysis. Rick likes to know that he's on the same page with new employees. He wants to know that you understand the industry the way he does; so, startbypreparing acoupleofslidesthatshowbrieflywheretheindustryisintermsofinnovationsandnew products and Flash's position in the industry.

(Antoinecollectshisnotepadandpenandrisestowalk.)

Theresa:Letmeknowifyouhaveanyquestions.

In preparation for starting his new position with Flash, Antoine has been reading up on the semiconductor industry. In hindsight, he is glad he did, as he thinks he will be able to draw more meaningful insights from his analyses with that background knowledge.

Among the things he has learned from his readings are that in the semiconductor industry, quick changesinconsumertastescancausedramaticchanges indemandfordifferenttypesofchipsets. For example,asconsumers increasinglyusedtheirphonesforactivities,suchastexting andcalendaring, correspondingactivities,suchasemailing,declined,andwiththathavecomedeclines indemandfor PCs and increases in demand for cellphones. Also, the increasing popularity of wearable computing devices,suchasfitness trackersandremoteaccesstechnology,likethosethatfacilitatesecuring one's house remotely or changing the temperature settings remotely have contributed to the increasing growth of the IoT sector. With these insights, Antoine felt more confident and comfortable about starting his variance analyses. He thought he now only must get the engineering perspective.

Antoine felt comfortable about the concept of Moore's Law, but he did not fully appreciate the implicationson themarket segments thatFlash servedand how themarket segmentsin turn affected demandforthevariousproducts.Hedecidedtospendtherestoftheafternoon todosomeresearch on the implications of Moore's Law on Flash's market segments, and by extension, its products and thencomeupwithquestionstoaskJohnandTheresa. Perhapstheycanintroducehimtoanengineer or two who can help him better comprehend the technical side of the products.

ThenextdayAntoinelearnedfromaseasonedengineerwhomanagedone ofFlash'sfabsthatthenewer marketsegmentstendtorequiremoreadvancedchipsetsthataremoredifficulttoproduce.4Asaresult, themanagertendsto assignmoreexperiencedengineerstoproducethemorecomplexchipsets.Still,ittakesa coupleof years,accordingtothe fabmanager,tohit thestandardquantitiestheyshouldbe using.

From his research that afternoon, Antoine discerned that one implication of Moore's Law is that directmaterial pricevariancesshouldlargelybefavorable,sincetransistorcostsarebecomingcheaper evenascomputingpowerincreases.HeconcludedMoore's Lawthenappears tobedrivingsuchnew marketsegmentsasIoTanddatawarehousing.Still,hethought itagoodideatodiscusshisthoughts with the engineers. He'd talk to Dave, the fab manager, before leaving for the day.

Laterthatday,AntoinepaysDaveavisitinhisfaboffice.

Dave:Howarethings goingAntoine?Crunchtimeforthenumbercrunchers,right?

Antoine:You got it! Dave, I'm trying to wrap my mind around this Moore's Law idea and the cost of makingchips.Ifthechipcostsaresupposedtobegoing downalmostexponentially,whydoourunit costs ever go up?

Dave:I suspect part of it is that for the newer and higher-end chips, such as SSDs, 5we need to use more experienced labor. For example, I tend to do quite a bit of the work on the fab floor myself for SSDs-and I'm not exactly cheap labor. (He said this with a smirk.)

Antoine:Gotit.Butthatisonthelaborsideoftheequation,whataboutthematerialcost?

Dave:Itisthesameidea.Managementwantstousecheaperlabor butthatkidjustoutofengineering schoolisnotgoing toalwaysrunhisprocessescorrectlyhisfirst fewyears.

Antoine:Igetit.Thankyou.CanIcallorstopbyifIhaveanymorequestions?

Dave6:Anytime!

Supposeyou are in Antoine's position, yoursecondjobaftergraduation.Howwouldyou go aboutexplainingthevariancesbetween Flash's2022planandtheactualperformanceinasclearly and concisely as possible?

BackgroundonInput Costs

Inrecentyears,therehasbeenamovetowardmorestableandfasterNANDmemoryasopposed to the more volatile DRAM. NAND memory is not lost when there is no power, unlike DRAM.NAND-based memory, however, requires more power to run and is more expensive. Within NAND, there are the Single Level Cell (SLC) and Multi Level Cell (MLC). While the MLC can store moredata, it is not as fast as the SLC, so the SLCs are more expensive. Solid State Drives (SSDs) are the nextgenerationofNANDchips,andbecausetheyhavenomechanicalparts, thereisnolatency in accessing data. More explanation of these technologies is in Figure 1. Types of labor, minimum qualifications, and costs appear in Figure 2.

Figure1.Description Material	of Direct MaterialInputs and Related Products8 Description	ApplicableProducts(MarketSegments)
DRAM	Dynamic Random Access Memory. Each bit of data is stored in a memory cell. It has mechanical parts, is volatile, and needs power to store data,and is comparatively cheap.	PCs,IoTs,Other
NAND-MLC	The name is derived from Logic; Not And (NAND). MLC stands for multi- level cell. With this, stored data is not lost when power is off. Storesmore data than NAND-SLC but is slower than NAND-SLC.	PCs,Servers
NAND-SLC	SLCstandsforsingle-levelcell.NAND-SLCdoesnotlosedatawhenpowerisoff.StoreslessdatathanNAND-MLCbutisfasterthanNAND-MLC	Servers,Other
NAND-SSD	SSD denotes solid-state drive. This component has no mechanical parts, making data access much faster. Too expensive to use in large applica-tions like servers.	IoTs,Other
NAND-eMLC(Enter-prise-MLC)	This represents the next generation of NAND MLC being developed for large applications like use in servers and with many writes per day.	Servers
SiliconWafer-(i.e.,200mmor300mmindiameter)	Wafers are round, flat, very thin smooth glass-like products that form thebasic building block for manufacturing microprocessors and chips. They are made from sand and are very good conductors of electricity, hence,their use in the industry.

Figure2.2022LaborIn LaborType	putData Description	Cost
FabTechnician/Engineer	Bachelor's degree or some college education and up to five years' experi-ence	$25/hour
SeniorFabEngineer	5to10yearsofexperienceorMastersdegreeinarelatedfield	$30/hour
Lead Fab Engineer	Morethan5yearsofexperienceasseniorfabengineer	$35/hour
FabManager	More than 7 years of fab experience and some operational or businessexperience	Fixedsalaryof$180,000

Figure3.Personalitiesin Theresa	theCase Treasurer
John	AssistantTreasurer
Antoine	SeniorFinancialAnalyst
Rick	ChiefFinancialOfficer
Dave	FabManager

MICROPROCESSORSANDCHIPSETMANUFACTURING

Siliconwafersarefabricatedincleansterilizedfactoriescalledfabs.Overtheyearsasmicroprocessorshave become smaller, their sensitivity has increased, necessitating a stricter manufacturing environment that in turn increases the cost of a fab. Today, fabs cost anywhere from $3 billion to $8 billion, depending on the application. Deviations from strict manufacturing standards, including the higheststandardsofcleanrooms,canresultinlowyieldsofmicroprocessorsperwafer,addingtothe cost of manufacturing.

Overhead

Flash's overhead costs have many components. To simplify overhead application, with the help of consultants, Flash determined that Direct Labor Hours (DLH) captured most accurately between 80% and 90% of its overhead costs. Therefore, Flash uses DLHs to apply both variable and fixed manufacturing overhead (VMOH and FMOH). Fringe benefits like healthcare, pension, and depreciationofequipmentareexamplesoffixedoverhead,whilevariableoverheadincludeselectricity andsupplieslikescrubsthatemployeesweartoworkincleanroomsandplasticandceramicmaterials for packaging dies that make up the microprocessors or chips

FabManagerSalary

Though fab manager salaries are fixed, the managers tend to work more directly on higher-end products. As a result, the portion of their time spent more directly on the products are treated as directlaborhours.Theremainderoftheirsalariesaretreatedasoverheadandareincorporatedinto estimating the overhead rate.

Abriefoverviewofthemicroprocessormanufacturingprocess,includingaprocessflowdiagramofthemanufacturing process (Figure 4) and related terminology, is in Appendix A.

BUSINESSBACKGROUND

Flash designs and manufactures advanced digital technology microprocessors and chipsets into platformsthataretargetedtodiverseapplicationsintheinformationtechnologyandcommunication industries.Foundedin1975bytwoengineerswhosawanopportunitytobeearlyplayersinthethen burgeoningcomputerindustry,Flashhasgrowntobeoneofthetopsuppliersofmicroprocessorand chipset-based products to the information technology (IT) and communications industries.

Product/EndUsers

Flashsellsitsplatforms(chipsetsandmicroprocessors)tooriginalequipmentmanufacturers(OEMs), original design manufacturers (ODMs), and industrial and communication equipment manufacturers.Flashclassifiesitsproductsintermsoftheendusermarketsegmentsinwhichitsplatformsareused. The top categories are:

1. Thespectrumofcomputing:tablets,cellphones,notebooks,desktops
2. IoT:retaildevices,wearables,suchassmartwatches,autoapplications,homesurveillance,andsoon.
3. Serversforspecializedapplicationssuchasdatacenters,networktrafficmanagement,security,anddatabase management systems.

Strategy

Likeothertopplayersintheindustry,FlashhasusedMoore'sLaw?asaguidetodriveinnovationand productivity.Putforthin1965byGordonMoore,oneofthefoundersofIntel,Moore'sLawpositsthat thenumberoftransistorsonachipsetwilldoubleapproximatelyeverytwoyears.Inotherwords,the sizeofachipsetwillbehalved,orproductivitydoubledeverytwoyearsorso,dramaticallyreducing thecostofcomputersandrelatedproductsanddrivinginnovation.AnimplicationofMoore'sLawis thatcomputersandrelatedproductshaveincreasinglybecomesmalleryetmorepowerfulovertime.

Cell phones and all the applications that are now capable of running on them, such asgeographic positioning systems (GPS) technology, smart watches, and driverless car technology, are all the result of a"naturalgrowth"alongthetrajectoryimplicitinMoore'sLaw.Withanexponentialgrowthoverseveral decades,however,Moore'sLawishittingitslimitsasitbecomesmoreandmorechallengingtosqueeze more transistors onto the same space, especially as more transistors mean increased power consumption. Simultaneously,increased internet connectivity has resulted in more data generatedfrom diverse sources,helpinglaunchcloudcomputingandanexplosionofdatacentersasagrowthopportunityforFlash.

Flash's long-term strategy assumes that computing devices, such as cell phones and watches, will increasingly become extensions of the person, and companies will require more efficient data management and analysis capabilities as consumers generate increasing amounts of data. Flash has incorporated these trends into its current five-year technology roadmap. Guided by Moore's Law, the technology industry can project upcoming user needs and applications and coordinate withchip equipment manufacturers and chip manufacturers to plan accordingly, thus helping improve efficiency and innovation. Despite the great effort the industry puts into forecasting demand, firms can be victims of sudden changes in consumer tastes, and disruptive advances in technologies can result in large deviations from plans. With the help of an intern, Antoine collected data from Flash's accounting system for his analyses. Those financial and operational information are in Appendix B.

Duetorapidobsolescence,Flashstrivesto?anddoes?sellallitsproductsinthesameperiod.That is, there is no beginning and ending inventory.

APPENDIXB.DATAFORVARIANCEANALYSIS

	IncomeStatementbyGeographyandMarketSegment($Millions)
				2022	2022Plan	2021	2020
US&Americas				$3,030.24	$2,840.66	$2,773.26	$2,492.65
EMEA				$2,479.29	$2,982.70	$2,357.27	$2,492.65
Asia				$8,264.30	$8,379.96	$8,735.78	$8,862.75
Total				$13,773.84	$14,203.32	$13,866.32	$13,848.05
	IncomeStatementbyMarketSegment($Millions)
				2022	2022Plan	2021	2020
PersonalComputing				$1,154.59	$1,469.33	$1,466.29	$1,420.44
IoT				$2,906.76	$2,768.00	$2,800.16	$3,156.30
Servers&DataWarehousing			$7,256.15	$6,817.19	$6,373.73	$5,875.49
Other				$2,456.33	$3,148.80	$3,226.14	$3,395.82
TotalRevenue				$13,773.84	$14,203.32	$13,866.32	$13,848.05
LessCOGS				$6,886.92	$7,101.66	$6,933.16	$7,024.03
GrossMargin				$6,886.92	$7,101.66	$6,933.16	$6,824.03
Selling,General&AdministrativeExpenses	$5,184.47	$5,255.23	$5,026.54	$5,044.84
NetIncome				$1,702.45	$1,846.43	$1,906.62	$1,779.18

			2022Revenues($Millions)
	Product	SellingPrice
	Costperunit	PerUnit	PC-related	IoT	Servers	Other	Total
DRAM	$4.19	$8.38	$855.06	$1,482.10	$0.00	$513.03	$2,850.18
NAND-MLC	$2.73	$5.45	$299.53	$0.00	$449.30	$0.00	$748.83
NAND-SLC	$3.33	$6.66	$0.00	$0.00	$1,076.31	$993.52	$2,069.83
NAND-SSD	$3.85	$7.70	$0.00	$1,424.67	$0.00	$949.78	$2,374.45
NAND-eMLC	$5.61	$11.22	$0.00	$0.00	$5,730.54	$0.00	$5,730.54
Total			$1,154.59	$2,906.76	$7,256.15	$2,456.33	$13,773.84

			2022PlanRevenues($Millions)
	Product	SellingPrice
	Costper unit	PerUnit	PC-related	IoT	Servers,etc.	Other	Total
DRAM	$4.41	$8.82	$1,129.11	$1,247.97	$0.00	$594.27	$2,971.35
NAND-MLC	$2.87	$5.74	$340.22	$0.00	$433.01	$0.00	$773.22
NAND-SLC	$3.70	$7.40	$0.00	$0.00	$1,010.28	$1,453.81	$2,464.09
NAND-SSD	$4.05	$8.10	$0.00	$1,520.03	$0.00	$1,100.71	$2,620.75
NAND-eMLC	$5.50	$11.00	$0.00	$0.00	$5,373.91	$0.00	$5,373.91
Total			$1,469.33	$2,768.00	$6,817.19	$3,148.80	$14,203.32

			2022Actual
	PC-related	IoT	Servers	Other	Total
DRAM	102.05	176.88	-	61.23	340.16
NAND-MLC	54.95	-	82.42	-	137.37
NAND-SLC	-	-	161.68	149.24	310.92
NAND-SSD	-	185.14	-	123.43	308.57
NAND-eMLC	-	-	510.74	-	510.74
Total	157.00	362.02	754.85	333.90	1,607.76

PC-related	IoT	Servers	Other	Total	%ofTotal
DRAM	128.02	141.49	-	67.38	336.89	21%
NAND-MLC	59.29	-	75.46	-	134.75	8%
NAND-SLC	-	-	136.58	196.55	333.13	21%
NAND-SSD	-	187.66	-	135.89	323.55	20%
NAND-eMLC	-	-	488.54	-	488.54	30%
Total	187.31	329.15	700.58	399.81	1,616.86	100%

Table4,PanelA:ActualAllocationsofDirectMaterialInputPurchases(MillionsofUnits)

			2022Actual
	Price/Unit	PC-related	IoT	Servers	Other	Total
DRAM	$1.05	105.11	182.19	-	63.07	350.36
NAND-MLC	$0.69	56.60	-	84.90	-	141.49
NAND-SLC	$0.84	-	-	158.45	146.26	304.70
NAND-SSD	$0.96	-	192.55	-	128.37	320.91
NAND-eMLC	$1.40	-	-	561.82	-	561.82
Total		161.71	374.74	805.16	337.69	1,679.29

Note:Throughoutthecase,partialunitsareleft"asis"onlytofacilitatemoreaccuratecalculationsandminimizeroundingerrors. Partialunitswouldnormallyberoundedupotherwise.Alldatainthistableareactual.

Table4,PanelB:ActualAllocationsofDirectmaterialInputstoProducts(i.e.,used,millionsofunits)

			2022Actual
	Price/Unit	PC-related	IoT	Servers	Other	Total
DRAM	$1.05	102.05	176.88	-	61.23	340.16
NAND-MLC	$0.69	54.95	-	82.42	-	137.37
NAND-SLC	$0.84	-	-	161.68	149.24	310.92
NAND-SSD	$0.96	-	185.14	-	123.43	308.57
NAND-eMLC	$1.40	-	-	510.74	-	510.74
Total		157.00	362.02	754.85	333.90	1,607.76

Note: All data in this table are "actual". One input (e.g., DRAM) per output (e.g., PC-related). Throughout the case, partial units are left "as is" only to facilitate more accurate calculations and minimize rounding errors. Partial units would normally be rounded up otherwise.

Table4,PanelC:AllocationsofDirectmaterialInputstoProducts(StandardQuantityAllowed)

			2022Plan
	Price/Unit	PC-related	IoT	Servers	Other	Total
DRAM	$1.10	103.09	175.09	-	59.94	338.11
NAND-MLC	$0.73	55.51	-	82.84	-	138.35
NAND-SLC	$0.94	-	-	162.50	146.10	308.60
NAND-SSD	$1.01	-	183.26	-	120.83	304.09
NAND-eMLC	$1.38	-	-	513.32	-	513.32
Total		158.60	358.35	758.66	326.87	1,602.47

			2022Actual
	FabEngr	SnrFabEngr	LeadFabEngr	FabMgr	Total
CostperHour	$26.25	$30.90	$35.00	$61.48
DRAMHrs	21.38	6.58	-	-	27.96
MLC+SLCHrs	20.48	-	6.57	-	27.05
SSDHrs	-	15.24	4.51	-	19.76
eMLCHrs	-	-	-	25.69	25.69
Total	41.86	21.83	11.08	25.69	100.45

		2022Plan
	FabEngr	SnrFabEngr	LeadFabEngr	FabMgr	Total
Costper Hour	$25.00	$30.00	$35.00	$58.00
DRAMHrs	20.36	6.79	-	-	27.14
MLC+SLC Hrs	19.32	-	6.44	-	25.76
SSDHrs	-	14.25	4.75	-	19.00
eMLCHrs	-	-	-	24.70	24.70
Total	39.68	21.03	11.19	24.70	96.60

Flash,Inc.:AComprehensiveVarianceAnalysisCase

We will work through how to calculate the sales, direct materials, and director labor variancesduringtheliveclass.Note:IrecommendfollowingalonginExcel.Havingthe calculations/reports available will be helpful as you prepare the post-class assignment. You are not required to turn anything in before the live class.

LIVECLASS:

1. Suppose you are in the role of Antoine Banks, a recent graduate of a top-ranked MBA program. He must explain the changes in profits to Rick, the CFO, and Theresa,theTreasurer.Preparesalesvariancesfor2022vs.planforallproducts using the information in the appendix of the case to facilitate the calculations where necessary.
2. Preparepriceandusagevariancevs.planforalltypesofdirectmaterialinputs for 2022 using the information in the appendix of the case.
3. Preparedirectlaborrateandefficiencyvariancesvs.planforalltypesoflaborfor 2022 using the information in the appendix of the case.

POST-CLASS:

Submitaprofessionalmemosummarizingthefindingsandrecommendationsfrom this analysis.

RecommendedDiscussionPoints:

• Brieflyexplainthetoptwoproductsandmarketsegmentsthatcontributedthe most (by price and volume) to the sales variances.
• Brieflyexplainhowthedirectmaterialanddirectlaborvarianceshavedeviated from the plan or budget.
• WhatarethetoptwocontributorstoeachtypeofDMandDL variance?
• Giventheinformationinthecasespeculateonwhatfactorsmightbedriving these variances.
• Briefly explain how all the variances you have calculated collectively help to explainFlash's2022performancevs.itsplanorbudget.Stateanyassumptions you make upfront.

AdditionalInformation

ChangesinIncomeStatementbyMarketSegmentand2022Common- Sized Income Statement

2022 Sales 2022 Vs. Plan 2022 Vs. 2021 2021 Vs. 2020 Personal Computing 8% -21% -21% 3% IOT 21% 5% 4% -11% Servers & Data Warehousing 53% 6% 14% 8% Other 18% -22% -24% -5% Total Revenue 100% -3% -1% 0% Less COGS 50% -3% 1% - 1% Gross Margin 50% -3% -1% 2% Selling, General & Administrative Expe 38% -1% 3% 0% Net Income 12% -8% -11% 7%