# Iterating the prioritization criteria
<a name="iterating-prioritization-criteria"></a>

Before you create migration wave plans, we recommend that you iterate the application prioritization criteria to pivot from pilot application selection to long-term wave planning. 

In earlier sections, we introduced a default prioritization criteria that would prioritize simple cloud-ready applications (see [Prioritizing applications](prioritization-and-migration-strategy.md#prioritizing-applications)). This was because in early stages we recommend starting with noncritical applications to refine migration processes and incorporate lessons learned. However, at this stage, and to create long-term plans, the order in which applications are migrated should be aligned to business drivers. Applying the new criteria will generate a new ranking of applications that will be a key input for wave planning.

Review the available data points from the application portfolio, and select the attributes that will determine application prioritization based on business drivers.

First, validate your business drivers (see [Business drivers and technical guiding principles](business-drivers-technical-guiding-principles.md)). Next, based on your business drivers, select the attributes that will help to prioritize applications for migration. 

The following table shows example prioritization criteria aligned to business drivers for innovation.



- **Operating system**
  - ****Possible values**:** AIX / ****Score (0-99)**:** 80
  - ****Possible values**:** Solaris / ****Score (0-99)**:** 80
  - ****Possible values**:** HP-UX / ****Score (0-99)**:** 80
  - ****Possible values**:** Mainframe / ****Score (0-99)**:** 70
  - ****Possible values**:** Windows / ****Score (0-99)**:** 50
  - ****Possible values**:** Linux / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** High (1x)

- **Business criticality**
  - ****Possible values**:** High / ****Score (0-99)**:** 60
  - ****Possible values**:** Medium / ****Score (0-99)**:** 40
  - ****Possible values**:** Low / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** High (1x)

- **Architecture**
  - ****Possible values**:** Tightly coupled / ****Score (0-99)**:** 60
  - ****Possible values**:** Loosely coupled / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** High (1x)

- **Operating model**
  - ****Possible values**:** Traditional - no CI/CD / ****Score (0-99)**:** 60
  - ****Possible values**:** Basic CI/CD / ****Score (0-99)**:** 40
  - ****Possible values**:** Full DevOps / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** Medium-high (0.8x)

- **Number of compute instances**
  - ****Possible values**:** 1-3 / ****Score (0-99)**:** 60
  - ****Possible values**:** 4-10 / ****Score (0-99)**:** 40
  - ****Possible values**:** 11 or more / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** Medium-high (0.8x)

- **Migration strategy**
  - ****Possible values**:** Refactor (re-architect) / ****Score (0-99)**:** 70
  - ****Possible values**:** Replatform / ****Score (0-99)**:** 40
  - ****Possible values**:** Repurchase / ****Score (0-99)**:** 30
  - ****Possible values**:** Rehost / ****Score (0-99)**:** 10
  - ****Importance or relevance multiplying factor**:** Medium (0.6x)



The following table shows example prioritization criteria aligned to business drivers for quick cost reduction.



- **Database product**
  - ****Possible values**:** Oracle / ****Score (0-99)**:** 70
  - ****Possible values**:** Microsoft SQL / ****Score (0-99)**:** 70
  - ****Possible values**:** Others / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** High (1x)

- **Operating system**
  - ****Possible values**:** Windows / ****Score (0-99)**:** 70
  - ****Possible values**:** Linux / ****Score (0-99)**:** 70
  - ****Possible values**:** Others / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** High (1x)

- **CPU utilization (average)**
  - ****Possible values**:** More than 36% / ****Score (0-99)**:** 60
  - ****Possible values**:** Less than 36% / ****Score (0-99)**:** 40
  - ****Importance or relevance multiplying factor**:** High (1x)

- **Number of compute instances**
  - ****Possible values**:** 11 or more / ****Score (0-99)**:** 60
  - ****Possible values**:** 4-10 / ****Score (0-99)**:** 40
  - ****Possible values**:** 1-3 / ****Score (0-99)**:** 20
  - ****Importance or relevance multiplying factor**:** Medium-high (0.8x)

- **Migration Strategy**
  - ****Possible values**:** Retire / ****Score (0-99)**:** 80
  - ****Possible values**:** Rehost / ****Score (0-99)**:** 70
  - ****Possible values**:** Replatform / ****Score (0-99)**:** 50
  - ****Possible values**:** Refactor (re-architect) / ****Score (0-99)**:** 10
  - ****Importance or relevance multiplying factor**:** Medium (0.6x)



Test the prioritization criteria and iterate until you generally agree with the output. It takes at least three or four iterations to obtain a baseline version.