:

Junior Media Buyer: Get Healthy and Get Paid -

Tuesday, June 23, 2020

Is Salesforce a Great Place to Work? -

Thursday, December 26, 2019

Is Apple a Great Place to Work? -

Wednesday, June 19, 2019

Do I want to work at Adobe? -

Wednesday, June 12, 2019

NVIDIA, why work here? -

Friday, June 7, 2019

Manager Instructional Technology at George Washington University -

Thursday, October 4, 2018

5 Highest Paying Business Development Manager Jobs in New York -

Monday, July 23, 2018

What kind of Business Development
Jobs are in Los Angeles?
-

Wednesday, May 9, 2018

QVC , On Air Program Host Job for 3rd Largest Ecommerce Company -

Monday, March 26, 2018

Facebook has over 1700 Jobs: Here is How to Get a Job at Facebook -

Wednesday, February 7, 2018

Nielsen Why You Want to Work at this Digital Transformation Organization -

Tuesday, February 6, 2018

Why You Want a Job at Twitter -

Monday, February 5, 2018

How fast is this Blockchain thing going to take over? -

Friday, February 2, 2018

Should You Work at HBO or Netflix? -

Wednesday, January 31, 2018

Why Working at Hearst is Much Better than Houghton Mifflin Harcourt -

Friday, January 26, 2018

What Will Making a VR Game While in Virtual Reality be like? -

Thursday, January 25, 2018

Why You Want to Work at Snapchat -

Wednesday, January 24, 2018

Why You Want to Get a Job at Vogue Magazine: -

Thursday, January 18, 2018

Is it Better to work at Buzzfeed or The New York Times? -

Friday, January 12, 2018

LeBook Business Development Job for Trend Setter -

Wednesday, December 6, 2017

Making Job Search Easier by Finding the Great Companies First

Find a
JOB
Title/Keywords Company Name
City, state or zip (optional)
 

Techcrunch Interviews: Piqed — Optimizing Beauty Products for Consumers

Techcrunch Interviews: Piqed - Optimizing Beauty Products for Consumers

Can data lead to a more beau­ti­ful you?  Helen and Annie from Piqued (pro­nounced like picked) think so.  They have cre­at­ed a prod­uct designed to help you find the best beau­ty prod­ucts for your skin and look.  Using a data dri­ven mod­el and infor­ma­tion from your past expe­ri­ences Piqed will help you choose the beau­ty prod­ucts that will work best for you.

The glob­al cos­met­ics mar­ket was $460 bil­lion in 2014 and, accord­ing to Hahn, their focus is on the pre­mi­um seg­ment of the busi­ness which she esti­mates at $50 bil­lion.

Avail­able now as an IOS app the ser­vice is expect­ed to have a more for­mal launch in the fall.

I spoke with Helen Hahn and Annie Peng from Piqed and learned why they start­ed this busi­ness,  the prob­lems they are solv­ing and about the busi­ness oppor­tu­ni­ty.

Roy: This is Roy Weiss­man from MediaJobs.com. Today we’re talk­ing with Helen Hahn and Annie Peng from Piqed.

 

Helen: We tried to solve a prob­lem that a lot of women face when choos­ing skin­care prod­ucts. You walk into a store and maybe from hearsay or a friend’s rec­om­men­da­tion, you pick up a prod­uct. You try it and then you real­ize that it does­n’t work for you, no mat­ter how well it might work for the per­son next to you. This process is very inef­fi­cient. It’s cost­ly and also it could be poten­tial­ly dan­ger­ous if you have an aller­gic reac­tion to the prod­uct that you’re try­ing out.

 

We want to elim­i­nate that prob­lem com­plete­ly by help­ing our users find prod­ucts that are tai­lored just to them. How we do this is we actu­al­ly use a data dri­ven mod­el in the back that looks at their past expe­ri­ences with dif­fer­ent prod­ucts, whether it’s worked for them before or things that did­n’t work for them are equal­ly good data as well. Using that data, we gen­er­ate tai­lored, per­son­al­ized rec­om­men­da­tions just for that user, much like how Net­flix can dis­cov­er movies that you might be inter­est­ed in.

 

Roy: But if I used a prod­uct, how do I have data? Peo­ple just know I used a prod­uct. It was­n’t good. I did­n’t like it. It seems very sub­jec­tive or very qual­i­ta­tive, not quan­ti­ta­tive.

 

Helen: That’s a great ques­tion. From the user’s input, it is the qual­i­ta­tive data that we’re look­ing for that actu­al­ly no one records because no retail­er will ask you, “How well did it work for you?” to give you a rec­om­men­da­tion lat­er. In the back, cou­pled with that qual­i­ta­tive data, the quan­ti­ta­tive data is the ingre­di­ents of the prod­ucts that are being rat­ed or being rec­om­mend­ed, just like traits of a movie. I keep using this exam­ple because I think it’s eas­i­er to under­stand. The traits of a movie that Net­flix might use to cat­e­go­rize it, will be then used to then gen­er­ate rec­om­men­da­tions for that user, if that makes sense.

 

Roy: If I used a prod­uct, like a mois­tur­iz­er for my face and I said, “Oh, it did­n’t seem to mois­tur­ize my face. My face was dry ten min­utes lat­er.” How would that turn into quan­ti­ta­tive data? What do I do if I want to use your ser­vice?

 

Helen: You just tell me … You will look for that prod­uct in the data­base in our app, and then you’ll tell me the rat­ing. We ask you how well did it work for you or if it caused any kind of bad reac­tions. We take that data cou­pled with that pro­duc­t’s ingre­di­ents. That’s the “quan­ti­ta­tive” part of that prod­uct. That’s what dri­ving the prod­uct to work or not work for you. Giv­en that, the col­lab­o­ra­tive fil­ter­ing tech­nique then runs on all those spe­cif­ic ingre­di­ents and match­es it with things that are sim­i­lar, or in the case where if you did not like this prod­uct, things that are dis­sim­i­lar.

 

Roy: How do you know which of the ingre­di­ents were good and which of the ingre­di­ents were bad?

 

Helen: That is the pow­er of the actu­al AI algo­rithm, which is I don’t need to know the spe­cif­ic ones. If there is enough data dri­ving the mod­el, it will take care of itself. Of course it’s weight­ed by con­cen­tra­tion of the ingre­di­ents. The list goes by con­cen­tra­tion, if you were to ever look at the pack­ag­ing of the prod­uct.

 

Roy: The algo­rithm looks at the con­cen­tra­tion of item A ver­sus item B and my issues, my feed­back on each item.

 

Helen: Exact­ly.

 

Roy: Based on that, it deter­mines a mod­el that says, “This prob­a­bly is not work­ing because of X,” kind of thing, and then it goes from there. Is there any­one else doing any­thing like this?

 

Helen: We looked very hard and thor­ough­ly before we start­ed the ven­ture and we do not believe there is any­one who is tak­ing this approach. The very pop­u­lar approach is using mar­ket­ing lan­guage. A prod­uct will say, “We are com­bat­ing wrin­kles,” or, “We are com­bat­ing acne,” and then that becomes the cat­e­go­riza­tion for that pro­duc­t’s func­tion­al­i­ty, even though it may or may not be any dif­fer­ent from a prod­uct that just claims to mois­tur­ize.

 

Roy: You said you’re using an AI algo­rithm. Where is this com­ing from? Who is writ­ing this secret algo­rithm?

 

Helen: I think Annie can take that one.

 

Annie: Helen’s doing all the back­end stuff. She’s doing all the algo­rithm cal­cu­la­tion stuff.

 

Helen: We build every­thing in house, design front-end, back-end. It’s Annie, me and anoth­er founder.

 

Roy: You guys wrote the soft­ware.

 

Helen: Yes, we did.

 

Roy: What is your back­ground?

 

Helen: I stud­ied engi­neer­ing and math in col­lege and I was a high-fre­quen­cy algo­rith­mic trad­er, so an elec­tron­ic trad­er in finance.

 

Annie: I have busi­ness and biol­o­gy back­ground, but I went Cod­ing Boot­camp and I’m learn­ing cod­ing right now as well. I’m respon­si­ble for front-end stuff.

 

Roy: Is this just an app or do I do it online?

 

Helen: For now it’s just an iOS app, but we’re launch­ing the web plat­form very soon. We plan to expand to Android of course as well.

 

Roy: What is the size? Do you have any idea of the size of the mar­ket in finan­cial dol­lars that you’re talk­ing about?

 

Helen: This is almost astro­nom­i­cal­ly hard to answer as a ques­tion because the mar­ket is so huge. The skin­care mar­ket itself is pro­ject­ed to be about $120 bil­lion world­wide USD and of that, $50 bil­lion are going to be the pre­mi­um mar­ket, so peo­ple who are look­ing for very spe­cif­ic things. That’s just skin­care prof­its as gen­er­al. For us, we can also look at the mar­ket­ing bud­get of each retail­er. That would apply to, for us, if we were to become lead gen­er­a­tion for them giv­en that we give good rec­om­men­da­tions to users who will prob­a­bly buy their prod­uct and then use it and retain it, that par­tic­u­lar mar­ket, I think each retail­er is some­where from 60 to 80% of their gross prof­its are com­plete­ly for mar­ket­ing. It’s a huge, huge mar­ket in bil­lions and bil­lions of dol­lars, even just a tiny sec­tor of that for if we can help our users would be real­ly ben­e­fi­cial for them and for our­selves.

 

Roy: In essence, your app is sell­ing every­one else’s prod­ucts.

 

Helen: We are gen­er­at­ing the leads for them. These are sol­id, sol­id leads, unlike, say, adver­tis­ing or hearsay.

 

Roy: If I’m using your prod­uct and I deter­mine that Shi­sei­do or an Estee Laud­er cos­met­ic, Clinque, what­ev­er, is the right item, I just go to the store and buy it. How does any­one even know that I dis­cov­ered that in your app?

 

Helen: We have a click through buy links to be of con­ve­nience for the user. Most like­ly, if you are using the app and you dis­cov­ered it, you want to, espe­cial­ly if you’re into online shop­ping, you want to pur­chase it right away. The click through is there to help you do that.

 

Roy: You’re look­ing to do deals with com­pa­nies, online mer­chants, what­ev­er, to have them pay you a com­mis­sion, like an affil­i­ate com­mis­sion, when you sell one of their prod­ucts.

 

Helen: That would be our basic mod­el, yes.

 

Roy: Have you guys raised any mon­ey for the busi­ness yet?

 

Helen: We’re still in the seed fund­ing stage. We’re actu­al­ly start­ing an accel­er­a­tor tomor­row, part­nered with Google. We’re going to start that process soon.

 

Roy: You’re in the Google accel­er­a­tor?

 

Helen: It’s the GSV Labs, Google Launch­pad Accel­er­a­tor pro­gram. It’s in Sil­i­con Val­ley. It actu­al­ly start­ed today, but we’re here today.

 

Roy: That’s excit­ing. When did you found the busi­ness?

 

Helen: Feb­ru­ary is when we launched the app.

 

Roy: Do you have any sense of how many users you have at this point?

 

Helen: We have about 500 sol­id users that we’ve retained. Those are sol­id, organ­ic users.

 

Roy: Have they been buy­ing prod­ucts?

 

Helen: For now, we just track the actu­al buy­ing link, so at the very end, we don’t know if they end up pur­chas­ing it, but we know that they’re going through the link to find the prod­uct. Since we’re not cur­rent­ly part­ner­ing with a retail­er, we just want to grow our user base and we want to grow our user data base. Before we feel like we have a sol­id han­dle on that, we don’t want to mon­e­tize it very quick­ly, if that makes sense.

 

Roy: Do you feel you have an edge? What’s the true bar­ri­er to entry with all these great devel­op­ers out there, what do you think the true bar­ri­er to entry is where you’re going to have a head start and it’ll be hard for some­body to take away your busi­ness.

 

Helen: I hate to say this, but most of the devel­op­ers that can do this kind of machine-learn­ing algo­rithm are men and unfor­tu­nate­ly, as we’ve also wit­nessed today, they don’t nec­es­sar­i­ly think of skin­care as their first point of entry. We believe our idea is there first and we can bet­ter improve it faster.

 

Roy: Where would you like to see this busi­ness in five years?

 

Helen: We’d love it if every­one can use the app to dis­cov­er their holy grail skin­care reg­i­men and nev­er have to try some­thing that they don’t like ever again.

 

Roy: What kind of per­son would be great for your com­pa­ny?

 

Helen: We love peo­ple who have a pas­sion for design and data. I think data is prob­a­bly more impor­tant because we can always lay­er a dif­fer­ent design on the prod­uct itself, but it’s the actu­al data and how you use it, how you ana­lyze it that’s dri­ving every­thing. Doing things more smart­ly with data is always a bet­ter way to go. That’s what we firm­ly believe in. Any­one who fits that is wel­come to look us up.

 

Roy: Thank you very much.

 

Print Friendly, PDF & Email
Leave A Comment

You must be logged in to post a comment.