Vivek Sub­ra­man­ian recently left Ora­cle, where he led prod­uct man­age­ment efforts for Siebel Mar­ket­ing and Ana­lyt­ics, to join Calm­Sea, a startup focused on help­ing online retail­ers con­duct near real-time opti­miza­tion of crit­i­cal mer­chan­dis­ing func­tions such as pric­ing, mark­downs, assort­ment, pro­mo­tions and inven­tory. Calm­Sea is a mem­ber of the Omni­ture Devel­oper Con­nec­tion com­mu­nity, and is cur­rently work­ing to inte­grate its science-based pre­dic­tive mer­chan­dis­ing solu­tion with trans­ac­tion data from Omniture’s Gen­e­sis APIs. The joint solu­tion promises to offer online retail­ers a pow­er­ful new analytics-based tool designed to help ensure that they always have the right mix of prod­ucts at the right price avail­able to customers.

Q: In the online mar­ket­ing indus­try we hear a lot about strate­gic issues related to 3 of the 4 P’s; Prod­uct, Pro­mo­tion, and Place­ment, but rarely ever do we hear about the most impor­tant of the 4 — Price! We all learned in Econ 101 that price is a crit­i­cal deter­mi­nant of demand-why do you think online mar­keters often ignore it?

A: Great ques­tion to start off! Online retail­ers, with a few excep­tions, have tra­di­tion­ally been price fol­low­ers rather than price set­ters. Pric­ing has not been a strate­gic action but a tac­ti­cal one instead with reac­tions to com­pe­ti­tion being the pri­mary dri­ver. This has led to price wars that have eaten into gross mar­gins. One of the key rea­sons for this is the lack of afford­able tools that can help make pric­ing deci­sions by look­ing not only at com­pet­i­tive data but also mar­ket­ing, cus­tomer, inven­tory and sup­ply chain data and do all this in near real-time. For exam­ple, while pric­ing a prod­uct you need to con­sider past sea­son­al­ity of sales and site traf­fic demand as well as whether the cur­rent inven­tory lev­els are over­stock or under-stock. This will lead to the most ‘opti­mal price’ and not sim­ply the low­est price. At Calm­Sea we call this process Con­tin­u­ous Mer­chan­dis­ing TM, and our mis­sion is to help retail­ers opti­mize every mer­chan­dis­ing deci­sion — from prod­uct assort­ment to pric­ing, pro­mo­tions and inven­tory — by com­bin­ing their data with com­pet­i­tive pric­ing infor­ma­tion and part­ner data like Omniture.

Q: Calm­Sea has built a solu­tion that can look at large vol­umes of data and deter­mine things like opti­mal price for a given prod­uct or opti­mal mark­down for an over­stocked item. How does the inte­gra­tion of online ana­lyt­ics data help you achieve this?

A: Online ana­lyt­ics data is an extremely crit­i­cal piece in mak­ing pric­ing deci­sions. Many tools today try to pre­dict demand based on his­tor­i­cal sales trans­ac­tions while ignor­ing the por­tion that did not con­vert into sales — what we call as latent demand. Online site ana­lyt­ics pro­vides valu­able con­tex­tual data around why the prospect did not con­vert, their pref­er­ences and where they came from, to com­ple­ment mer­chan­dis­ing and sup­ply chain data. For exam­ple, it will help pin­point if the prob­lem is with a lack of site traf­fic demand gen­er­a­tion or with con­ver­sions due to a pric­ing prob­lem. We can thus drive holis­tic pric­ing deci­sions that will drive more con­ver­sions thereby improv­ing the bottom-line for the retailer.

Q: What spe­cific online ana­lyt­ics data from Omni­ture are you plan­ning to inte­grate with your system?

A: At Calm­Sea, we are inter­ested in all pieces of data that help our cus­tomers opti­mize mer­chan­dis­ing deci­sions. We believe that Omni­ture can cover the entire mar­ket­ing side of the equa­tion. For starters we will pull met­rics like basic site traf­fic and con­ver­sion data to help us opti­mize the price point, while loca­tion spe­cific site traf­fic data helps us drive local­ized pric­ing for cus­tomers who want that. We com­bine the same data with inven­tory and his­tor­i­cal sales data to pre­dict future over-stock or under-stock sit­u­a­tions. In addi­tion, we can lever­age the click-stream path analy­sis to dis­cern prod­uct affin­ity that in turn can be used to drive bun­dled pro­mo­tional offers and prod­uct assort­ment plan­ning for the next sea­son. We think we have only scratched the sur­face, there’s so much more we can do with things like mar­ket­ing cam­paign effec­tive­ness metrics.

Q: Once a cus­tomer is up and run­ning, how long does it take for them to get pre­dic­tive feed­back on pric­ing and assort­ment from your system?

A: It typ­i­cally takes us a week or so to get a mid-sized retailer up and run­ning. Once they are up they get vis­i­bil­ity into their busi­ness met­rics imme­di­ately. They can also start man­ag­ing and opti­miz­ing prices and pro­mo­tions imme­di­ately. Our sci­ence plat­form is a self-learning adap­tive plat­form and the deci­sions will improve with time.

Q: For the typ­i­cal online retailer, what do you expect the impact from deploy­ing Calm­Sea will be in terms of mar­gin improve­ment, top-line rev­enue growth, etc?

A: This kind of solu­tion has proven ROI for the brick-and-mortar retail space. Many of us in fact come from com­pa­nies that suc­cess­fully built such soft­ware prod­ucts and sold to cus­tomers like Safe­way, Best­Buy and Wal­mart. Such solu­tions have proven up to 15% improve­ment in rev­enue and mar­gin. The online world, we believe, pro­vides many more oppor­tu­ni­ties to improve on this because of newer mar­ket­ing chan­nels and more effi­cient mer­chan­dis­ing actions like real-time price changes. Our work with our ini­tial beta cus­tomers tells us that we will see sig­nif­i­cantly bet­ter num­bers in online retail.

Q: While offline retail is still the major­ity of dol­lar trans­ac­tions, the online chan­nel is becom­ing a sig­nif­i­cant part of the retail expe­ri­ence. Is there an oppor­tu­nity for online ana­lyt­ics to influ­ence offline merchandising?

A: I think you hit upon an impor­tant ongo­ing macro change in the mar­ket. Today 40% of all offline retail sales (USD 937 bil­lion) are influ­enced in some way or the other by the online chan­nel. By 2014 this num­ber will be 54%. So online retail expands beyond just what you sell through the online chan­nel. For multi-channel retail­ers there is a huge oppor­tu­nity to opti­mize across the two chan­nels using the strengths of one to com­pen­sate for the weak­nesses of the other. For exam­ple, one of our apparel cus­tomers uses the low-cost strength of the online chan­nel to test new prod­ucts for con­sumer pref­er­ences by loca­tion and then stock appro­pri­ate prod­ucts in phys­i­cal stores. We believe Omni­ture and Calm­Sea are at the heart of this macro-movement.

Q: On a tech­ni­cal note, you’ve built your appli­ca­tion in Adobe Flex. It’s a beau­ti­ful, clean interface…but beyond aes­thet­ics, why did you choose Flex as your devel­op­ment platform?

A: We have been extremely sat­is­fied with the Flex frame­work and the com­mu­nity around it. Aside from the aes­thet­ics I would say that the pro­duc­tiv­ity it pro­vides is phe­nom­e­nal after the ini­tial learn­ing curve. The pre-built com­po­nents and open-source work atop the Flex frame­work have hugely reduced our time-to-market.

Q: Besides ana­lyt­ics, what else is on your inte­gra­tion roadmap with Omniture?

A: Beyond online ana­lyt­ics we see two major areas of inte­gra­tion in the near-term. First, we plan to inte­grate with Omniture’s Test & Tar­get mod­ule to enable easy A/B test­ing of price changes. Sec­ond, we plan to inte­grate with Omniture’s mer­chan­dis­ing solu­tion to pro­vide a com­plete solu­tion that includes ana­lyt­i­cal tools as well as exe­cu­tion tools.

Rudi Shumpert
Rudi Shumpert

Great read! It's nice to hear more about integrations with other Adobe products. -Rudi