Failure  mode  effect  analysis    (FMEA)  
 
This  is  a  tool  developed  in  the  field  of  quality  management  which  systematically  
tries  to  imagine  different  ways  in  which  a  system  might  fail  and  then  puts  in  
place  contingency  plans  for  avoiding  it  or  minimizing  the  risks  s  associated  with  
it.    Originally  developed  in  the  1950s  as  military  systems  became  more  complex  
the  technique  has  been  refined  and  extended  and  now  forms  an  important  part  
of  project  planning  for  a  variety  of  innovations  in  fields  as  diverse  as  
construction  and  healthcare.    It  is  sometimes  called  FMECA  where  the  ‘C’  refers  
to  criticality  –  of  the  many  ways  in  which  the  system  could  fail  which  are  the  
critical  ones?    Identifying  these  and  providing  contingency  plans  for  dealing  with  
them  is  the  primary  design  task.  
 
FMEA  is  essentially  a  brainstorming  exercise  in  which  teams  with  particular  
knowledge  and  perspectives  explore  in  systematic  fashion  the  possibilities  of  a  
system  failing.    Typically  there  are  five  stages  in  an  FMEA:  
 
• Identifying  failure  modes  –  what  could  go  wrong?  
• Identifying  root  causes  of  these  failures  –  why  would  the  failure  happen?  
• Identifying  failure  effects  –  what  would  the  consequences  of  these  failures  
be?  
• Identifying  and  prioritizing  criticality  –  which  failure  modes  are  most  
significant?  
• Developing  ways  to  deal  with  these  failure  modes  –  designing  them  out,  
designing  fail-­‐safe  options,  identifying  contingency  plans,  etc.  
 
FMEA  is  a  powerful  design  tool  when  creating  a  new  product,  service  or  process  
but  it  can  also  provide  valuable  systematic  knowledge  around  continuous  
improvement  as  the  organization  acquires  experience  of  the  innovation  in  use.  
 
The  process  of  carrying  out  an  FMEA  is  as  follows.  
 
1 Assemble  a  team  representing  different  perspectives  of  the  project  –  
for  example  marketing  (to  represent  customer/user  needs),  
engineering,  manufacturing,  software,  etc.    They  need  to  be  diverse  
enough  to  think  through  the  different  angles  of  the  proposed  new  
product/service  or  process.  
2 Agree  on  the  boundaries  of  the  FMEA  –  essentially  define  what  is  the  
process/product  or  service  which  will  be  examined.    In  processes  this  
can  often  be  done  using  a  flow  chart,  in  products  a  product  map  and  in  
services  a  user  journey.  

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
 3 Using  a  simple  matrix,  put  the  stages  of  the  process,  the  elements  of  
the  product  or  the  steps  on  the  service  journey  in  the  first  column.    
Let’s  take  the  example  of  a  customer  in  a  restaurant  ordering  a  meal.    
The  service  journey  might  look  as  follows:  
 
Stages  in  journey      
Seat  customer      
Give  menu  and  return      
in  five  minutes  to  take  
order  
Take  order  and  pass  to      
kitchen  
Collect  order  from      
kitchen  after  food  is  
ready  
Serve  customer      
Return  after  time      
delay  to  clear  plates  
and  prepare  bill  
Present  bill  and  take      
payment  
Present  change,      
receipt,  etc  and  escort  
customer  out  
 
There  are  many  additional  stages  we  could  add  but  let’s  keep  it  simple  for  this  
example.  
 
4 For  each  of  these  then  fill  in  a  second  column  which  looks  at  all  the  
possible  ways  failure  could  take  place  –  failure  modes.  
 
Stages  in  journey    
Seat  customer   Fail  to  seat  customer  
Give  menu  and  return   Fail  to  give  menu  
in  five  minutes  to  take  Fail  to  return  at  all  
order   Fail  to  return  in  
specified  time  period  
Take  order  and  pass  to   Fail  to  pass  to  kitchen  
kitchen  
Collect  order  from   Fail  to  collect  order  
kitchen  after  food  is  
ready  
Serve  customer   Fail  to  serve  customer  
Return  after  time   Fail  to  return  
delay  to  clear  plates   Return  to  clear  plates  
and  prepare  bill   but  fail  to  prepare  bill  
Present  bill  and  take   Fail  to  present  bill  
payment   Fail  to  take  payment  
Present  change,   Fail  to  present  
receipt,  etc  and  escort   change/receipt  
customer  out   Fail  to  escort  customer  
out  
Joe  Tidd  and  John  Bessant  
http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
  
Once  again  there  are  many  other  ways  in  which  things  can  go  wrong  –  part  of  the  
brainstorming  with  the  team  is  to  imagine  worst-­‐case  scenarios  where  anything  
and  everything  that  can  go  wrong  is  recorded.    Role-­‐play  and  simulation  are  
useful  tools  in  this  stage.  
 
5 For  each  failure  mode  add  a  third  column  which  looks  at  the  
consequences  of  that  failure  happening.  
Stages  in  journey   Failure  mode   Consequences  
 
Seat  customer   Fail  to  seat  customer   Customer  gets  
annoyed,  may  leave  
Give  menu  and   Fail  to  give  menu   Customer  gets  
return  in  five     annoyed,  may  leave  
minutes  to  take   Fail  to  return  at  all   Table  is  occupied  but  
order     non-­‐productive  =  
Fail  to  return  in   opportunity  cost  
specified  time  period  
Take  order  and  pass   Fail  to  pass  to   Customer  gets  
to  kitchen   kitchen   annoyed,  may  leave  
 
Collect  order  from   Fail  to  collect  order   Food  gets  cold  
kitchen  after  food  is   Customer  gets  
ready   annoyed,  may  leave  
Wasted  food,  wasted  
effort  in  kitchen,  
opportunity  cost  
since  table  is  
occupied  but  not  
earning  money  
 
Serve  customer   Fail  to  serve   As  above  
customer  
Return  after  time   Fail  to  return   Customer  may  get  
delay  to  clear  plates     annoyed  and  
and  prepare  bill     complain  or  even  
  leave  
  Food  has  been  
  consumed  so  all  the  
  waste  above  plus  
  customer  may  refuse  
  to  pay  
   
Return  to  clear   Customer  may  walk  
plates  but  fail  to   out  without  paying  
prepare  bill    
 
Present  bill  and   Fail  to  present  bill   As  above  
take  payment   Fail  to  take  payment    
 
Present  change,   Fail  to  present   As  above  plus  no  
receipt,  etc  and   change/receipt   tip!!!!  

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
escort  customer  out   Fail  to  escort    
customer  out  
 
 
6 In  the  next  column  put  a  figure  to  indicate  the  severity  or  impact  of  
that  failure  happening,  typically  on  a  scale  of  1  (=  minor  impact)  to  10  
(=  major  impact)  
Stages  in  journey   Failure  mode   Consequences   Severity  of  impact  
 
Seat  customer   Fail  to  seat  customer   Customer  gets   5  
annoyed,  may  leave  
Give  menu  and   Fail  to  give  menu   Customer  gets   5  
return  in  five     annoyed,  may  leave  
minutes  to  take   Fail  to  return  at  all   Table  is  occupied  but  
order     non-­‐productive  =  
Fail  to  return  in   opportunity  cost  
specified  time  period  
Etc   Etc   Etc   Etc  
       
 
7 In  the  next  column  list  the  potential  causes  (ideally  the  root  cause)  
which  might  cause  that  failure  to  happen.    (tools  like  a  fishbone  
diagram  can  be  helpful  here)  
Stages  in   Failure  mode   Consequences   Severity  of   Potential  
journey     impact   causes  
Seat  customer   Fail  to  seat   Customer  gets   5   Overcrowded  
customer   annoyed,  may   restaurant.  
leave   No  clear  
queuing  or  
priority  
system  
Too  many  
customers  
for  each  
waiter  to  
deal  with  
Give  menu  and   Fail  to  give   Customer  gets   5   Lack  of  
return  in  five   menu   annoyed,  may   training  
minutes  to  take     leave   Too  busy  –  
order   Fail  to  return  at   Table  is  occupied   result  of  
all   but  non-­‐ wrong  waiter  
  productive  =   to  customer  
Fail  to  return  in   opportunity  cost   ratio  
specified  time  
period  
Etc.   Etc   Etc   Etc   Etc  
         
 
8 For  each  failure  add  another  column  which  indicates  the  likely  
frequency  of  occurrence  –  again  using  a  simple  scale.    For  example  1  
(=  very  rarely)  through  to  10  (=  inevitable)  
Stages  in   Failure   Consequences   Severity  of   Potential   Likely  

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
journey   mode     impact   causes   frequency  
of  
occurrence  
Seat   Fail  to  seat   Customer  gets   5   Overcrowded   7  
customer   customer   annoyed,  may   restaurant.  
leave   No  clear  
queuing  or  
priority  
system  
Too  many  
customers  
for  each  
waiter  to  
deal  with  
Give  menu   Fail  to  give   Customer  gets   5   Lack  of   5  
and  return   menu   annoyed,  may   training  
in  five     leave   Too  busy  –  
minutes  to   Fail  to   Table  is   result  of  
take  order   return  at  all   occupied  but   wrong  waiter  
  non-­‐productive   to  customer  
Fail  to   =  opportunity   ratio  
return  in   cost  
specified  
time  period  
Etc.   Etc   Etc   Etc   Etc   Etc  
           
 
9 For  each  failure  mode  add  another  column  in  which  you  identify  how  
that  fault  might  be  detected  (process  control)  and  if  it  occurs  how  to  
deal  with  it  to  minimize  the  impact.  
Stages   Failure   Consequenc Severit Potential   Likely   Process  
in   mode   es   y  of   causes   frequenc controls/  
journe   impact   y  of   recovery  
y   occurren actions  
ce  
Seat   Fail  to   Customer   5   Overcrowd 7   Monitor  
custom seat   gets   ed   number  of  
er   custom annoyed,   restaurant.   customers  and  
er   may  leave   No  clear   install  queuing  
queuing  or   system.    Head  
priority   waiter  to  
system   oversee  
Too  many   booking  and  
customers   balance  
for  each   allocation  to  
waiter  to   tables  
deal  with   Reduce  
waiter/custo
mer  ratio  
Give   Fail  to   Customer   5   Lack  of   5   Etc  
menu   give   gets   training    
and   menu   annoyed,   Too  busy  –  
return     may  leave   result  of  

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
in  five   Fail  to   Table  is   wrong  
minutes   return   occupied  but   waiter  to  
to  take   at  all   non-­‐ customer  
order     productive  =   ratio  
Fail  to   opportunity  
return   cost  
in  
specifie
d  time  
period  
Etc.   Etc   Etc   Etc   Etc   Etc    
             
Stages   Failure   Consequence Severit Potential   Likely    
in   mode   s   y  of   causes   frequency  
journey     impact   of  
occurrenc
e  
Seat   Fail  to   Customer   5   Overcrowd 7    
custom seat   gets   ed  
er   custom annoyed,   restaurant.  
er   may  leave   No  clear  
queuing  or  
priority  
system  
Too  many  
customers  
for  each  
waiter  to  
deal  with  
Give   Fail  to   Customer   5   Lack  of   5    
menu   give   gets   training  
and   menu   annoyed,   Too  busy  –  
return     may  leave   result  of  
in  five   Fail  to   Table  is   wrong  
minutes   return   occupied  but   waiter  to  
to  take   at  all   non-­‐ customer  
order     productive  =   ratio  
Fail  to   opportunity  
return   cost  
in  
specifie
d  time  
period  
Etc.   Etc   Etc   Etc   Etc   Etc    
             
             
 
There  are  other  columns  which  can  be  added  –  for  example  how  easy  is  the  error  
to  detect?    How  early  in  the  process  can  it  be  detected?  Will  the  controls  be  
sufficient  to  detect  it?  And  so  on  –  the  point  is  to  think  through  thoroughly,  using  
the  different  perspectives  in  the  team,  all  the  ways  in  which  things  might  go  
wrong  and  then  to  build  in  ways  of  minimizing  or  eliminating  these  problems.  

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd  
  
The  above  example  is  for  a  service  journey  but  we  can  see  the  same  pattern  
working  in  process  innovation.    In  product  innovation  this  FMEA  activity  can  
lead  to  redesign  and  improvement  of  features  –  and  also  help  to  generate  the  
'troubleshooting’  section  which  often  appears  in  the  user  manual  accompanying  
the  product.  
 
For  more  on  FMEA  see  http://asq.org/learn-­‐about-­‐quality/process-­‐analysis-­‐
tools/overview/fmea.html  
Other  related  tools  include  cause-­‐and-­‐effect  diagrams  (fishbone  charts)  and  
poke-­‐yoke  mistake  proofing  
 
 
 
 
 
 
 

Joe  Tidd  and  John  Bessant  

http://www.innovation-­‐portal.info/  
John  Wiley  and  Sons  Ltd