Editor’s Note: The Interview is done for TechNode by Timothy Coghlan, an expert on the fashion industry in China and writes about the China fashion business in his blog maosuit.

With a database of over 6 million street style photos taken across China’s biggest cities and through big data analysis, Chinese lifestyle focused social network P1 is now able to track and predict fashion trends as the develop across China.

Screenshots of P1 Mobile App

Screenshots of P1 Mobile App

Starting this Saturday, P1 will host its Great Style Leap exhibition in Beijing’s Taikooli shopping district. The Style Leap exhibition tells the story of China’s rapidly evolving fashion scene driven by urban trendsetters.Prior to the Exhibition’s opening, Technode spoke to P1’s Marketing Director Svante Jerling about P1’s use of technology and big data to dissect fashion trends, and their upcoming exhibition.

Please tell us a bit about how P1 got started?

We stared P1 in 2007 because we saw the business opportunity to create a social network targeted at wealthy urban Chinese. These people lead significantly different lifestyles from other sectors of society and to them ‘dinner’ means a high-end hotel, and ‘education’ means Harvard or high school abroad. We realized these people needed somewhere to hang out online and so they became our target users.

Initially, our customer and user acquisition model was to take photos of these wealthy urbanites out on the town, at clubs and night events and then share the photos online. By doing this we acquired the content (photos) and users simultaneously as people would sign up for P1 to see and share the photos of their night out etc. [This was before phone cameras became ubiquitous].

But then we realized that our target users were all hanging out in shopping malls too. So we started taking street style photos and began acquiring content and users in shopping environments too. To maintain exclusivity we kept P1 an invitation only social network.

What changes have you seen take place amongst your users over the last six years?

Since 2007 until now we have taken over 6 million photographs, mostly in first tier cities – Beijing, Guangzhou, Shanghai, Shenzhen and also a bit in Chengdu, Chongqing, Dalian and Qingdao. With over one million (verified) different people in the photos we realized we have this massive street style database that represents the development of Chinese fashion trends.

As a whole, the photos act as a historical record of the Chinese street style across the last six years, which given China’s speed of development, can be likened to decades of fashion evolution in other countries.  Its fascinating to see that within such a short time frame we have seen the paradigm shift of thoughts and values amongst young people expressed through what they wear.

When we started taking photos in 2007-08, everyone was choosing what to wear with ‘status seeking’ in mind, where success was determined by how many luxury brands and distinguishable logos you had on display. Now, six years later, across the photos we can see a kaleidoscope of individual expression through fashion and there’s a diverse range of fashion tribes emanating across the country.

Fashion for the young urbanites no longer just about status and gone are the external standards of what society thinks defines success and what is ‘acceptable’ to wear (meaning you buy the most expensive thing you can afford). Now in 2013 we see Chinese adopting to fashion in terms of self-actualization.

How are you using big data able to track these trends through P1?

As most people we photograph are happy to give us their contact details, over six years P1 has built a massive data bank about which hipster, in which city, in which mall was wearing what colors. This, in addition to tracking the trends in the photos real-time as they are starting, gives us very powerful tool to measuring trends and ‘cool’.

These days everyone goes on about Key Opinion Leaders (KOLs). Well, with fashion we have found that often the KOLs probably aren’t the people actually initiating a trend. They may be an early adopter and help drive the trend, but they aren’t starting them – they are in fact just copying someone else – the initiators.

We created our own algorithms that at first just tracked colors of clothes in the photos and then we added another dimension to track objects such as bags, dogs and fixed gear bicycles, plus record the frequency of them occurring etc. This is where we start to make sense of the big data and apply our analytics.

The analysis will feed us information and give scenarios that predict the upcoming trends. So, for example if these 50 key individuals have been seen wearing item X or Y, then we can determine what chance the trend has to get big and take China by storm.

So through the photos and our database of users, we can trace trends back to their roots and identify the key initiators who are starting these movements. Then we can decipher, the conditions needed and chances of the trend really going big.

Does this give P1 and the power to influence trends and how is this affecting traditional fashion media?

P1 is in a position to act as a catalyst for trends, yet that’s not specifically our aim.

We don’t just focus on initiators who start the trends either. At P1 we focus on people we think are cool and then put them in the spotlight through our online and mobile platform, and it that way we have influence.

With this we also start to encroach into the domain of traditional printed media. Our algorithms can analyze magazines in exactly the same way as we analyze our street photos. When you put the two together, we can determine which fashion magazines have the most influence to drive trends according to by what items they feature in the magazines. This has huge potential to make print media more efficient and in-turn will influence advertisers spend etc.

How do you monetize your platform and this database of street style photos and analytics?

Our platform and affluent user base gives us multiple ways to work and collaborate with our brand partners and offer special services.  P1’s advantage has always been a large (wealthy) user base and our brand clients like the fact that we pinpoint people who are naturally cool and we aren’t trying to stage what is cool or deliberately force trends on our users.  Our database also gives us valuable information on users including: contact details, their color preferences, shopping habits and if they ride a fixed-gear bike etc. which is powerful for marketers.

Recently during Shanghai Fashion Week we collaborated with Starbucks on some product placement into our photos. Every 10 images or so featured someone holding a Starbucks Cup, or there was an (often blurry) Starbucks image in the background. In this way, the product placement was subtle and not so in your face like so much other advertising in China.

Brands acknowledge that capturing street style images can be more relevant than produced fashion campaigns because they are real and their target market consumers can relate to them better. However, this also makes it a very delicate balance with product placement because once it feels faked it backfires.

Therefore, we also need to be cautious and work with advertisers in a sustainable way and not let marketers abuse our platform to the point where it becomes played out. By ‘sustainable’ I mean not following the “I’m just going to plastering my image everywhere” philosophy that is pretty standard in China. Going back to our collaboration with Starbucks, the products were placed in a subtle way that didn’t bombard users.

What challenges do you have to taking the big data analysis to even more precise trend tracking and forecasting?

Altogether we use a combination of image analytics and algorithms to data mine the photos. Then a small percentage of the data must be entered manually by someone who can identify and tag the specific items and brands onto the photos.

The next level of analysis of the photos would be to create algorithms that can accurately identify brands and logos etc. Achieving this is mostly a technology issue, its not hard per se, just complex and so we need a lot of resources over time to solve this it.

To partially solve this issue, we have also introduced a tagging function into our App so users can tag items themselves and the content becomes more interactive and user generated.

Tell us about the P1 Great Leap Forward Style Exhibition that opens this Saturday

We decided to do this exhibition now because our data is finally sufficient enough to explain, qualify and prove all the trends we have seen and uncovered through our big data analysis. Therefore, the Exhibition will be interesting for fashionistas and techies alike.

The exhibition is a celebration of the development of individuality, which is something new in China. Chinese culture dictates that kids have to follow their parent’s wishes on what hobbies they pursue, what they study and who they date. Parents influence many aspects of young Chinese people’s their lives, except perhaps what they wear.  So clothes are their first taste of freedom and step towards choosing what they want for themselves.

We also want to show the world what Chinese fashion style. For the first time, China is moving from being the country where fashion is produced on mass to a country where global fashion styles and trends can originate. While a lot of the world thinks there aren’t many fashionable people here, the truth is that China could be an international style influencer in the next few years.