Market Segmentation with Qlik Set Evaluation and Qlik Set Operations

On this publish, we’ll evaluate two elusive strategies inside Qlik by which key enterprise questions may be addressed: Qlik Set Evaluation and Qlik Set Operations.

A standard enterprise goal is to broaden product gross sales or decide strategic effectiveness.  These issues typically take a type like one of many following questions and are requested with a watch towards historic efficiency.

• Which of my present clients bought my product?
• Which of my present purchasers are benefitting from my packages?

Qlik supplies an array of instruments to assist within the solutions to those questions.  We’ll use Qlik Set Evaluation to determine clients with particular traits or behaviors after which mix this with Qlik Set Operations to additional perceive the place we’d count on alternatives.

Qlik Set Evaluation

Our pattern knowledge set is a listing of fictitious clients and their orders.  We all know their geographic particulars and their order historical past.  From right here we will start to glean some historic tendencies and goal conduct, geographic or different attribute knowledge from which to determine extra gross sales alternatives.

Let’s start by figuring out these clients buying bikes.  Utilizing Qlik Set Evaluation we will determine these clients who’ve bought bikes up to now.  A method to do that is the next:

COUNT( { $ <PRODUCTLINE={"Bikes"}> } Distinct CUSTOMERNAME)

Within the desk beneath we see the shopper’s identify, a rely of shoppers and a rely of shoppers who’ve bought bikes.

Negating this, we’d then look forward to finding these clients NOT buying bikes.

COUNT({$<PRODUCTLINE-={"Bikes"}>} Distinct CUSTOMERNAME)

We see the two^nd and three^rd measure columns above will not be mutually unique.  Why is that this? 

What’s being recognized within the set are the ORDERS somewhat than the CUSTOMERS and whereas that is equal for the primary case, it’s clearly not for its negation within the second case. 

A simpler technique to realize this and retain the power to successfully determine the complimentary set is to make use of the P() and E() capabilities offered by Qlik for this function.

As a substitute of:

COUNT( { $ <PRODUCTLINE={"Bikes"}> } Distinct CUSTOMERNAME)

We use:

COUNT({$<CUSTOMERNAME=P({<PRODUCTLINE={"Bikes"}>})>}Distinct CUSTOMERNAME)

That is learn as ‘Which clients have EVER bought bikes’ the place P() signifies Doable.

To realize the complimentary set of these clients who’ve NEVER bought bikes [where E() indicates Excluded] we will do one of many following:

                COUNT({$<CUSTOMERNAME=E({<PRODUCTLINE={“Bikes”}>})>}Distinct CUSTOMERNAME)

– OR –

COUNT({$<CUSTOMERNAME-=P({<PRODUCTLINE={“Bikes”}>})>}Distinct CUSTOMERNAME)

We are able to now observe that for each buyer they both HAVE or HAVE NOT bought bikes.  (Word – as written, the Set Evaluation will retain context of any dimensional alternatives because of the $ notation).  As affirmation of this truth, we will see that the sum of the 2 teams (49 + 43) sum to the entire (92).

Qlik Set Operations

Because it stands, this may be helpful, nevertheless the strategies’ worth is amplified when mixed with different units by way of Qlik Set Operations.

COUNT({$
                <CUSTOMERNAME=P({<PRODUCTLINE={"Bikes"}>})>
    *
<CUSTOMERNAME=P({<PRODUCTLINE={"Planes"}>})> 
    } Distinct CUSTOMERNAME)

The Motorbike set ingredient is multiplied (*) with the Planes set ingredient to present us the intersection of those two units.  On this case, we’ve these clients who’ve EVER bought each Bikes AND Planes.  We are able to then rapidly manipulate the units to reply which ever questions we’d wish to pose.

Which clients have EVER bought bikes, however NEVER bought Planes?

COUNT({$
                <CUSTOMERNAME=P({<PRODUCTLINE={"Bikes"}>})>
    *
<CUSTOMERNAME=E({<PRODUCTLINE={"Planes"}>})>    
    } Distinct CUSTOMERNAME)

Alternatively:

COUNT({$
                <CUSTOMERNAME=P({<PRODUCTLINE={"Bikes"}>})>
    -
<CUSTOMERNAME=P({<PRODUCTLINE={"Planes"}>})>    
    } Distinct CUSTOMERNAME)

Qlik Set Operations Abstract

Combining Qlik Set Evaluation and Qlik Set Operations

If, as an alternative of in search of easy attribute identifiers, we want to perceive behavioral thresholds, i.e., Gross sales above $175k, we will leverage search in a extra superior Qlik Set Evaluation.

SUM({$<CUSTOMERNAME=P({<CUSTOMERNAME={"=SUM(SALES)>=175000"}>})>} SALES)

This may be additional altered and mixed by way of Qlik Set Evaluation Features P() and E() and Qlik Set Operations (* and -) to determine a really particular subset of shoppers for potential evaluation.

These clients…

SUM( {$
                // by no means having over 175k in gross sales (see E() exclude operate beneath)
                <CUSTOMERNAME=E({<CUSTOMERNAME={"=SUM(SALES)>=175000"}>})>
     *
// who've ever bought Planes (see P() potential operate beneath, * operator above)
    <CUSTOMERNAME=P({<PRODUCTLINE={"Planes"}>})>
     -
//however will not be positioned in USA or Australia (see subtraction operator above)
    <CUSTOMERNAME=P({<COUNTRY={"USA","Australia"}>})>
    } SALES)

See the ‘Mixed’ column beneath for the gross sales of the desired set of shoppers.

We now have the power to ask and reply questions which may goal subsets of shoppers primarily based on any attribute or conduct and which may be simply and reliably manipulated with out prolonged or advanced enhancing.