狠狠撸

狠狠撸Share a Scribd company logo

Using Lifecycle Scores for Marketing Optimisation

?
?
Carmen Mardiros
Carmen Mardiros

The document discusses using lifecycle scores to optimize marketing. It proposes calculating a "discovery score" to aggregate various customer behaviors on websites, like interacting with categories and products, into a single metric. This score represents the propensity of a marketing channel to move customers to the next stage of the purchase process. It describes a two-step process: 1) Assign a value or conversion rate to each behavior. 2) Calculate an overall discovery score by weighting each behavior by its ratio of visits from that channel. This score could then be used to evaluate channels, campaigns, landing pages and more to understand what is effectively moving customers along in their decision process.

1 of 15
Downloaded 23 times
Using Lifecycle Scores for Marketing Optimisation Carmen Mardiros @carmenmardiros Friday, 28 March 14
@carmenmardiros Customer’s cognitive decision process, not YOUR marketing funnel. Post-decision evaluation Need Recognition Option Evaluation Shortlist Interest (research) Decision (... Some loopbacks and optional steps, but some form of it always exists ) Friday, 28 March 14
@carmenmardiros Where your path crosses that of the customer... Post-decision evaluation Need Recognition Option Evaluation Shortlist Interest (research) Decision ... there’s just one goal - Move them to NEXT stage Friday, 28 March 14
@carmenmardiros Need Recognition Option Evaluation Shortlist Interest (research) Decision See category pages Actively interact with category pages See product pages Actively interact with products Discovery behaviours The One Goal Find a contender product, to buy now -- or later. (success measure = “Add to basket” ) Friday, 28 March 14
@carmenmardiros See category pages Actively interact with category pages See product pages Actively interact with products Engage with curated product lists Read pre-purchase help, tools and features Add to wishlist .... Intuitively, valuable behaviours but many and varied... Can we aggregate them all in a single “discovery score”? Organic PPC 30 350 4 7 21 7 3 13 99 460 7 34 1 1 = Propensity of channel to deliver valuable discovery visits. Friday, 28 March 14
@carmenmardiros Step 1. How much is each discovery behaviour worth? Users that interact with category pages Out of theses, users that ALSO add to basket A B Discovery score = B/A * 100 (... or the conversion rate for this type of discovery behaviour) Friday, 28 March 14
@carmenmardiros Step 1. How much is each discovery behaviour worth? Smoothen out ?uctuations by averaging over weekly cohorts Caveat: Would need to split by other dimensions to ?nd the true “value” of a behaviour. Tedious work until API supports visitor-level segments. ( Better yet, use median or con?dence ranges ) Friday, 28 March 14
@carmenmardiros Step 1. How much is each discovery behaviour worth? Behaviours closer to “Add to basket” tend to have higher discovery scores ( duh! ). Channels that drive a mix of high-score visits are better at moving people to NEXT stage in the decision process. Friday, 28 March 14
@carmenmardiros Step 2. Calculating discovery scores when channels bring a messy mix of visits score1 * ratio1 +score2 * ratio2 .... Discovery score = Why this formula? It evens the playing ?eld (e.g. channels driving more lower-score visits ~ channels driving fewer higher-score visits) Friday, 28 March 14
Use cases: - Validate intended response from campaigns and messaging - Uncover true channel purpose - Identify time-wasters (high engagement, wrong kind of engagement) - Identify marketing opportunities (low volume, high discovery score) Friday, 28 March 14
Not just for channels... Functional overlay for landing page groups to understand ?ows. (see Gary Angel - Functionalism) Friday, 28 March 14
Not just for channels... Function of content seen and interacted with. Friday, 28 March 14
Progression through decision process Not just for channels... Friday, 28 March 14
@carmenmardiros Yet to do, cross-reference with Multi Channel Funnels Friday, 28 March 14
Thank You Carmen Mardiros @carmenmardiros Friday, 28 March 14

More Related Content

Using Lifecycle Scores for Marketing Optimisation

  • 1. Using Lifecycle Scores for Marketing Optimisation Carmen Mardiros @carmenmardiros Friday, 28 March 14
  • 2. @carmenmardiros Customer’s cognitive decision process, not YOUR marketing funnel. Post-decision evaluation Need Recognition Option Evaluation Shortlist Interest (research) Decision (... Some loopbacks and optional steps, but some form of it always exists ) Friday, 28 March 14
  • 3. @carmenmardiros Where your path crosses that of the customer... Post-decision evaluation Need Recognition Option Evaluation Shortlist Interest (research) Decision ... there’s just one goal - Move them to NEXT stage Friday, 28 March 14
  • 4. @carmenmardiros Need Recognition Option Evaluation Shortlist Interest (research) Decision See category pages Actively interact with category pages See product pages Actively interact with products Discovery behaviours The One Goal Find a contender product, to buy now -- or later. (success measure = “Add to basket” ) Friday, 28 March 14
  • 5. @carmenmardiros See category pages Actively interact with category pages See product pages Actively interact with products Engage with curated product lists Read pre-purchase help, tools and features Add to wishlist .... Intuitively, valuable behaviours but many and varied... Can we aggregate them all in a single “discovery score”? Organic PPC 30 350 4 7 21 7 3 13 99 460 7 34 1 1 = Propensity of channel to deliver valuable discovery visits. Friday, 28 March 14
  • 6. @carmenmardiros Step 1. How much is each discovery behaviour worth? Users that interact with category pages Out of theses, users that ALSO add to basket A B Discovery score = B/A * 100 (... or the conversion rate for this type of discovery behaviour) Friday, 28 March 14
  • 7. @carmenmardiros Step 1. How much is each discovery behaviour worth? Smoothen out ?uctuations by averaging over weekly cohorts Caveat: Would need to split by other dimensions to ?nd the true “value” of a behaviour. Tedious work until API supports visitor-level segments. ( Better yet, use median or con?dence ranges ) Friday, 28 March 14
  • 8. @carmenmardiros Step 1. How much is each discovery behaviour worth? Behaviours closer to “Add to basket” tend to have higher discovery scores ( duh! ). Channels that drive a mix of high-score visits are better at moving people to NEXT stage in the decision process. Friday, 28 March 14
  • 9. @carmenmardiros Step 2. Calculating discovery scores when channels bring a messy mix of visits score1 * ratio1 +score2 * ratio2 .... Discovery score = Why this formula? It evens the playing ?eld (e.g. channels driving more lower-score visits ~ channels driving fewer higher-score visits) Friday, 28 March 14
  • 10. Use cases: - Validate intended response from campaigns and messaging - Uncover true channel purpose - Identify time-wasters (high engagement, wrong kind of engagement) - Identify marketing opportunities (low volume, high discovery score) Friday, 28 March 14
  • 11. Not just for channels... Functional overlay for landing page groups to understand ?ows. (see Gary Angel - Functionalism) Friday, 28 March 14
  • 12. Not just for channels... Function of content seen and interacted with. Friday, 28 March 14
  • 13. Progression through decision process Not just for channels... Friday, 28 March 14
  • 14. @carmenmardiros Yet to do, cross-reference with Multi Channel Funnels Friday, 28 March 14