PIWIK web data analysis(FUN88)

Web customer journey

• Customer journey contain so many interactions, if we only see the deposit report that difficultly tall us why the deposit decresed on July 05 2018?

  

We can use two different data source: BOSS backoffice data(Summary data) and PIWIK(Log data).

1. Summary data, we can choose some columns that may cause the deposit decreased:

• EX: I choosed four columns (Winning %, Bonus Applied Amount, Registered Member Count, Distinct Member Login Count) and find the correlaion between Approved Deposit Amount:

• From the correlation ‘Registered Member Count’ and ‘Distinct Member Login Count’ got more higher value, that means they are more effected on ‘Deposit’

• And then? It seems a commom sense, the more user login that will cause the deposit increse. The problem is ‘How to let our user login or become our members?’

2. PIWIK data(or GA data)

#Referral Data analysis

• Where our visit come from? ex: Visitor from referral 20180701

• What do these referrals preform?

If we can get the member code in PIWIK then we can combine the PIWIK with BOSS to discover the performance. Now we don’t have that, but we can use correlation between referral visitors and deposit amount to find out which referrals are valued?

• From the table we can see the higher correlation of referrals almost are the direct referrals (https://www.avskjed.com/), but still have some referrals with potential:

  

  http://www.tianhong56.com/xb.php

• On the otherside, if the correlation become negative that means users come from these referrals have low deposit amount:

  

  http://www.soq9.com/