This question reminded us of a very important quote i.e. “Success is 99% failures”.

It’s true that, often times the suggested optimizations doesn’t return the expected returns but, that doesn’t mean, the brands and retailers sit quietly and wait for the disaster to happen. The only way to survive in the online world is to constantly optimize based on change in the visitor behaviour. We have seen multiple optimization projects in the past which didn’t result in any significant returns, and caused lot of frustration among the ecommerce group and decline in the team credibility. Here are some of the common mistakes performed by the ecommerce team while executing the optimization projects:

• Most of the findings from the analyst are not properly verified and discussed with the ecommerce team. The team rush towards an optimization plan without verifying the source of the issue and impact on the quarterly and yearly business goals.
• The ecommerce team doesn’t spend enough time coming up with the alternatives to fix the issue. Often times, they restrict themselves to max. 1 to 2 alternatives, which are usually not accompanied with a ROI Model.
• Still, majority of the online players in the industry don’t believe in testing and targeting approach. In the absence of which, team wastes ridiculous amount of resources without any significant returns.
• Once the changes are published to production, there is always an expectation of immediate results. For some reason, every client I have worked in the past expects the results in first 1-2 weeks, and if it’s not there, they assume the project to be a failure.
• Every client I worked in the past asked for the benchmark data but, if you ask them to benchmark their own data before starting any optimization project, they either forgets it or don’t feel a need for it or, sometimes don’t collect it properly.
• Some optimization projects involves, changes in the reporting business rules which results in either discontinuity of trend or, change in the average conversion numbers. In both cases, the amount of discomfort it causes to the team is enormous. The reason behind this discomfort is the poor planning in regards to rules creation and publishing.

If being an ecommerce player, you have performed some of the above mentioned mistakes; here are some tips for your ecommerce team to create a result-oriented Optimization plan:

• Verify the Source and Measure the Impact:Every new finding from your web analyst should be verified with a microscopic eye. Here are 3 questions you should ask your web analyst:
  • Why is it a problem?
  • When it started happening?
  • What if, we don’t act on this problem?
• Unless you get the satisfactory answers to the above questions, do not rush towards the optimization plan. Here are some of the reasons why you shouldn’t rush:
  • Sometimes, the problem is seasonal or, caused by some promotional offer on the website but, do not have any impact on the quarterly and yearly business goals.
  • Sometimes, the problem is caused by incorrect business rules implemented against the reports.
  • Sometimes, the problem is not actually a problem but, lack of understanding about site structure.
  • Sometimes, the problem is caused by sudden change in the visitor behaviour after major site upgrade.
  • Sometimes, the problem is caused by recessionary forces.
  • Sometimes, the problem is caused by cut down in the market spend
  Once the problem is verified, the next step is to measure the impact on key conversion metrics and visitor behaviour. If the % of impact is very minimal and can be ignored in the short term, do not waste your precious resources on it. But, if the impact is significant and could deviates the brand from its quarterly or yearly goals; invest your precious resources on it.
• Benchmark the Current State: Once the problem is verified and impact analysis is performed, the next step is to benchmark the current state of the business by noting down the data against key conversion metrics and traffic trends.
• Create a ROI-Focused Optimization Plan: Once the data is benchmarked, it’s a time to create an optimization plan with following data elements:
  • Problem Definition
    • Description
    • Root Cause
    • Impact
  • Solution
    • Solution Description
    • Impact on Key Conversion Metrics
    • Alternatives (Atleast 3)
  • ROI Analysis
    • Anticipated Project Cost for Best Alternative
    • Anticipated Project Cost for Other Alternatives
    • Breakeven Time for Best Alternative
    • Breakeven Time for Other Alternatives
    • Expected ROI from Best Alternative
    • Expected ROI from Other Alternatives
    • Dependencies
    • Risk Areas
• Create a Testing & Targeting Approach: Over the last 5 years, Testing and Targeting are the two most talked about areas in the ecommerce industry but still, there are very few companies who have actually invested in this area. The reasons are:
  • Higher Cost of Ownership
  • Lack of Subject Knowledge
  • Sensitivity in regards to Launch Timelines
  • Scepticism in regards to higher Project Cost

  Irrespective of the fact whether, you are a new entrant in the ecommerce market or a seasonal ecommerce player, it is very important to use some kind of a test and target solution to validate your optimization approach. There are various solutions floating in the market from big vendors like Google, Coremetrics, Omniture etc. such as, Google Web Optimizer (FREE), SiteSpect and Omniture Test & Target.

  If you are a new player in the market with a limited budget and not proficient in this area then, we would highly recommend going for Google Web Optimizer because, it’s FREE, Intuitive and Economical to Maintain but, if you are a seasonal player looking for advanced capabilities and have decent budget then, either go for SiteSpect or Test & Target.

  While creating an Optimization Plan, make sure you are including a separate section for Testing and Targeting. These 2 solutions would allow you to test multiple alternatives simultaneously across different visitor segments, and target the best alternative to the entire site population once the statistically significant data is collected. On an average it takes 4-6 weeks to collect statistically significant data.

• Repeat the Optimization Cycle: With this, we conclude this post and as usual, we are looking forward for your valuable feedback and criticism.