Home | About us | FAQ / Questions | Software | ERP Implementation toolkit | Site map | Free checklists: ABCD WCM | Business Excellence Books | contact us


Best of breed (BOB) versus fully integrated system (FIS) which is best? by Phil Robinson

Best of breed (BOB) means using a specific software program or package for each specific application or requirement. To share information between the applications, the information is either printed out from one package and manually input to the next or the packages are linked either by the vendor or using a third party "middleware" package, sometimes called Enterprise Integration Technology (EIT), to provide varying degrees of integration. EIT generally requires a considerable amount of IT time to set up and maintain.

Fully Integrated Software (FIS) is a software package with a number of integrated modules to cover a range of functions and requirements. There are generally a number of databases automatically linked by key fields such as the part number or customer's name.

In a manufacturing company the principle fully integrated software is Enterprise Resource Planning (ERP) or variations such as Advance Planning Systems (APS) which have similar core functionality plus some extras. There are some applications required by manufacturing companies that ERP packages cannot handle, such as Computer Aided Design (CAD). Some functions ERP packages do handle but not generally as well as the best of breed packages, such as sales forecasting and customer contact management. Some functions ERP packages handle well, such as master production scheduling, material requirements planning, capacity requirements planning, inventory control, engineering data control (the bill of material in particular), purchasing and finance. In addition some integrated manufacturing packages have additional features such as finite capacity scheduling, payroll, planned maintenance and so on. First we need to look at what functionality a manufacturing company typically requires from its software.

Which software applications?
The table on the right shows the typical applications required by a manufacturing company. Software vendors might disagree but few people with experience in manufacturing planning would seriously disagree with the following table which shows where Best of Breed (BOB) is applicable and where Enterprise Resource Planning (ERP) software could be considered instead of the best of breed alternative.

Where there are question marks in the table, this indicates that some integrated manufacturing systems can do this task and some cannot or do not do it well so your choice depends on which package you are using or considering.

Visual planning systems
For a manufacturing company with a very small number of staff, say less than 10, the best planning system is a visual planning board where the jobs are represented by cards with the details written on the card. Time usually runs along the top, with the stages or equipment represented by the rows. The jobs can be moved about on the board as the work progresses. Production planning meetings are held around the board. If you want to take a "backup", you photograph the board! Fully integrated software is just not applicable although some specific best of breed applications may be applicable, accounts and contact management software for example.

Let me now deal with using spreadsheets for planning. The problem with spreadsheets is that they are generally owned by just one person. Spreadsheets are very satisfactory for small companies (less than 50 people typically) but, as the company grows, the necessity to copy information from one system to another is administratively slow, expensive and inevitably leads to mistakes. So for small companies with a relatively simple product, planning with spreadsheets is fine.

As a company grows it outlives spreadsheets and there are benefits of fully integrated software. If a company has implemented a planning system but still has information on spreadsheets that could and should be on the integrated system, this is a symptom that the planning system was either not properly implemented or is not being used properly. The spreadsheets undermine the integrity of the planning system for everyone.

To summarise so far, for small companies with, say, less than 50 people selective use of best of breed software is a clear winner. The main exception is where the company in question is a small part of a larger corporation with the need to exchange or combine data on a regular basis. It is with larger companies that the best of breed versus integrated software such as ERP is a more difficult decision.

ERP versus Best of Breed
There are 3 starting points that will affect the argument. You could start with little or no software, some software or you may have a full function package installed but are only using it for a single purpose such as finance or purchasing. I will deal with partial implementations below but start with the more straightforward case where a company has no planning software at present excluding spreadsheets so could, and probably would, use all the relevant functionality of an ERP package. I will compare ERP with best of breed solutions on 4 aspects

  • The time, cost and resources for implementation
  • Degree of fit
  • Simplicity of use
  • Implementation management

Time, cost and resources
It is generally agreed, both by consultants and companies who have implemented ERP systems, that to successfully implement an ERP package a company must review all its current processes to align them to the best way to work using the package. For example, with a manual system you can "get away" with poor data accuracy as people will spot obvious mistakes and because they know which data is uncertain they can make appropriate corrections. Not only do computer systems not allow for incorrect data but they compound data errors because each error affects all dependant and related data on all the integrated modules. A culture of data accuracy is an essential requirement for an ERP implementation to be successful. It is also generally agreed that, with the possible exception of finance, all ERP modules have to be implemented at the same time (big bang) as each module depends on the others. For these reasons the implementation of an ERP package requires a considerable company resource over a period of, typically, 12 to 18 months.

Implementing best of breed packages can be spread over a longer time period. Each package may only affect one department so can be implemented at a time when that department is available. On the other hand each best of breed package will have its own way of working. Anyone who has to work on more than one of the packages has to learn a number of completely different sets of instructions and whilst each may be simpler than the ERP equivalent, the total amount of learning will be considerably more. Each package will also come with its own set of constraints and problems where workarounds are required.

When negotiating for an ERP package, the customer has the advantage that the purchase is large so there is a better chance of cutting a deal not just on the software but also on on-going support. With the best of breed packages the ticket price may be lower but there is little chance of negotiating. As a general rule best of breed software will give good value for money on the specific job it does but will not come with the bells and whistles that an ERP package can have with its larger source of data.

In summary, if you can get the resources to implement an ERP package the overall cost and time required will generally be less than implementing the equivalent best of breed packages. Best of breed packages give you the option to spread the implementation over a longer time scale but will take more time overall and may actually cost more in the end.

Degree of fit
Because a best of breed package can be purchased to fit a particular need, they are almost bound to be the best fit. Every ERP package has its strengths and weaknesses; an ERP package that has good financials may be poor on process manufacture and so on.

The more expensive ERP packages (often called tier 1) have countered this problem by building in a huge amount of flexibility. Flexibility however also brings complexity and complexity means lots of consultants spending lots of time configuring the product to your exact needs which in turn means extra cost and time. To counter the cost and time required to implement these highly flexible packages, the implementation consultants have produced pre-configured "templates" for specific applications such as a food industry template or a textile template, but these templates reduce the flexibility of the package. With an integrated package you cannot have both simplicity and a good fit.

On the degree of fit question, best of breed packages win not just because they can be bought to meet an exact requirement but because they can usually be used "out of the box".

Simplicity of use
Best of breed software, as you would expect, is easy to use, otherwise it would not be considered the best of breed! ERP software can be easy to use but most are not. Part of the reason ERP software is harder to use is that when you enter a piece of information you are making the information available for a number of different modules. Thus a bill of material must be formatted to suit design, purchasing, manufacturing, planning and finance. This multi-use is a problem but does have a payback for ERP packages. Whilst, for instance, receiving goods into stock, may be more time consuming with an ERP package than for a stand alone accounts package, that information is now also available for planning, purchasing, stock control and manufacturing without any more data entry.

Some companies attempt to integrate best of breed packages but however much your IT guys say integration will cost you should multiply by 10! Occasionally integration can work particularly when the applications have been designed to talk to the other application. However, even if two applications can be made to transfer data reliably 100% of the time (99.9% is not good enough) problems arise if one of the applications is modified or upgraded when data can be lost with disastrous consequences.

Simplicity of use is therefore a balance. Task by task best of breed generally scores but from an overall company perspective ERP is the winner in most cases. The bigger the company, the more the ERP packages scores over best of breed because of the volume of data, the need for speed and the sheer geography of larger companies.

Project management
When it comes to project management ERP is a clear winner. With only one supplier to deal with, life is so much easier. Anyone who has implemented multiple packages that have any level of overlap will tell you that when there are problems each supplier will blame the others. Equally there is one company doing the training and support.

Upgrades are a necessary evil. Complex ERP packages do not have a good upgrade history. No-one wants to be the first to upgrade unless there is a serious problem that the upgrade theoretically fixes. The trick is to wait a while, talk to the user group or watch the bulletin board until the dust settles on the new release and then upgrade. At least with one package you have one point of contact if things go wrong.

With best of breed there will be a continual stream of upgrades. The more integration and data extraction that your IT department has done for you, the more work there will be at every upgrade.

To summarise, best of breed wins on degree of fit but a fully integrated package wins on time, cost, resources, simplicity and project management. The best of breed approach does allow you to spread the cost and resources required for implementation so is a short term fix.

Replacing a current ERP package
I am sure that many people reading this article will have an ERP system installed but do not feel they are getting the benefits they expected from it. The system has become more of a burden than an asset. If, due to your size or the nature of your business, you have concluded that best of breed is the answer then you should abandon the integrated package and implement best of breed packages. If a fully integrated system appears to be the right solution for you, your next job is to re-look at your current system.

The differences between ERP packages is not as much as their sales people want you to believe. The basic functionality is the same - add orders and material, subtract production and sales. Whilst the add-on are quite useful there is seldom a big enough difference to make it worth throwing away your current package and starting from scratch with a new one. The exceptions to this rule is if the software vendor has either gone out of business, has withdrawn support for the package or the package is so badly supported that it is unusable.

I have been involved with companies in the process of replacing a system which a similar company is busy installing. I also know that the success or failure of the implementations has little to do with the software. If you are not currently using your ERP package successfully, you will need independent outside help to re-implement, just trying harder is never going to work. It is a fact that most successful ERP implementations had the help of independent outside consultants who had experience of successful implementations.

Partial ERP implementations
To finish this report I will look at the situation when a company has concluded that a fully integrated package is right for them but does not want or need to implement a whole ERP package. If a company has, for instance, a financial package they are happy with, should they dump this and use the integrated package? The answer to this question is a definite yes. If a fully integrated package is right for the company, the people using stand alone systems must be trained in the use of the integrated system so that all the stand alone systems, databases and spreadsheets can be dumped. Not only will islands of data reduce the advantages of the integrated system but it will also undermine its integrity.

Another situation when a company does not need a full ERP implementation is when a package is either already implemented but not fully used or where only some of the modules have been implemented. In my experience for anyone but the smallest companies, the benefits of implementing the remaining modules is the most cost effective project the company is ever likely to have. There are 2 reasons for this.

ERP packages when properly implemented are very cost effective. When a chunk of the cost and some of the implementation work has already been completed, finishing the job is much cheaper and quicker than starting from scratch and so has an even better payback.

The differences between ERP packages, as mentioned above, is not enough to make it worth throwing away a partly implemented package and starting from scratch with a new one. The only exception to this rule is, as before, if the software vendor has either gone out of business or has withdrawn support for the package or it is so badly supported that it is unusable.

Limited resources
The final scenario is where a company just does not want to or cannot implement a full ERP package for resource reasons, money or time. In this case the company has to look into the future. If the balance of probability is that full ERP functionality will never be required, the best of breed is clearly the best solution. If the balance of probability is that the company will need all or most of the facilities that an ERP package will provide in the next 5 years then the decision should be to make the best of a bad job and at least implement the module(s) needed of a fully integrated system with a plan to implement the rest as soon as possible.


bobvfis2I started writing this article with a fairly open mind about the advantages of best of breed versus fully implemented systems. As I have researched each aspect I have come to the firm conclusion that for larger companies, say over 50 people, in most cases the core functionality of an ERP system is best served by an ERP system rather than a collection of best of breed packages. ERP systems do require considerable resources to implement properly but it is even better to implement an ERP system module by module than implement a collection of best of breed packages. The place best of breed scores is in the degree of fit but this is more than offset by the other factors.

There are many applications such as CAD, customer contact management, sales forecasting and many human resource functions where best of breed is the best solution with managed manual interfaces to an ERP system. As ERP suppliers strive to extend their functionality so these peripheral applications may increasingly be effectively carried out within the integrated system.

For larger companies that have an ERP system partially implemented or implemented badly, completing the implementation or re-implementing will be very cost effective.


Author : Phil Robinson

Published by BPIC - the business performance improvement consultancy

24, Hove Park Villas, Hove BN3 6HG, England - tel : +44 (0)1273 708561

Url : http://www.bpic.co.uk

Copyright Phil Robinson

All rights reserved. Reproduction for profit strictly forbidden. The material may be copied for personal and company internal use provided this resource paragraph is included.