At Effi­cient Fron­tier, we love to tease out pat­terns and use data to inform deci­sions. One pat­tern we’ve noticed is that the Aver­age Order Value (AOV), or the aver­age dol­lar amount spent for each cus­tomer order, tends to drop for retail­ers dur­ing the peak Christ­mas shop­ping sea­son. One the­ory to explain this unlikely phe­nom­e­non is that con­sumers who pur­chase gifts for oth­ers tend to buy less expen­sive items than if they were shop­ping for them­selves. We can also sur­mise that cer­tain peo­ple only become cus­tomers dur­ing cer­tain sea­sons, so the mix of cus­tomers changes.

The graph below shows the AOV trend for two groups of clients. Group A con­tains retail­ers that skew toward roman­tic and sen­ti­men­tal items com­monly gifted dur­ing Christ­mas or Valentine’s Day. Group B con­tains retail­ers that tend to sell more prac­ti­cal wares that are not typ­i­cally con­sid­ered roman­tic. We can see from the graph that both groups dip in AOV dur­ing the Christ­mas shop­ping sea­son. How­ever, Group A also dips to mid-December lev­els begin­ning around the third week of Jan­u­ary. Group B also takes a slight dip begin­ning on the 1st of Feb­ru­ary, just before Valentine’s Day.

Graph 1

That’s not to say that order vol­ume doesn’t play a role. It most cer­tainly does, though the vol­ume trends are not quite the same.

Graph 2
Besides the obvi­ous fact that not all retail­ers have the same sea­son­al­ity trends, we can see some sim­i­lar­i­ties that arise across the board. So why does any of this matter?

Whereas Christ­mas has a well-defined hol­i­day shop­ping “sea­son” (begin­ning on Black Fri­day), other hol­i­days, such as Valentine’s Day, don’t have such clearly defined shop­ping sea­sons. A change in AOV can indi­cate a change in the mix of cus­tomers. To cap­i­tal­ize on this, retail­ers can tai­lor ad copy towards cus­tomers who are seek­ing to pur­chase gifts.

Also, since only Group B showed a spike dur­ing Thanks­giv­ing when there were a lot of deals, this sug­gests that retail­ers should exper­i­ment with NOT dis­count­ing roman­tic goods by much.

In addi­tion to ad copy changes, cam­paigns need to be set up to cap­ture the increase in demand. This means dynam­i­cally chang­ing bud­gets and bids with the appro­pri­ate spend mix on high and low AOV cam­paigns. All of this could help max­i­mize con­ver­sions from cus­tomers who are plan­ning to buy gifts dur­ing the hol­i­day season.

Based on this data, we also rec­om­mend that retailers:

  • Use his­toric data to antic­i­pate demand, and pace bud­gets accordingly
  • Exper­i­ment with pric­ing on less time-sensitive but essen­tial retail goods, which show a high degree of elas­tic­ity. “Roman­tic” goods are time-sensitive, and do not usu­ally cap­ture addi­tional demand from low pric­ing dur­ing the high season
  • Align your cam­paigns in a cross-channel fash­ion. Use Face­book for upper-funnel opti­miza­tion to gen­er­ate demand early on in your hol­i­day cam­paign, and get aggres­sive on search to cap­ture lower-funnel intent late in the campaign

Ulti­mately, it all boils down to pre­dic­tive mod­el­ing and proac­tive account man­age­ment. In order to max­i­mize your ROI, you need best-of-breed adap­tive learn­ing algo­rithms to place the right bids at the right times. You also need a world-class team to pro­vide human insights and pro­vide the right cre­ative elements.

Joe Romero

Busi­ness Analyst