The idea Archer Chiang turned into a business in 2016 was simple: an online platform where users could order personalized gifts for each other. A serenade from a guitarist, for example, or a made-to-order fireworks display. Today, Chiang’s company, Giftpack, still offers a modified version of that service, but is quickly expanding into a more lucrative corner of the gift-giving world: AI-driven corporate gifting at scale.
Most startup founders agree that you should expect to pivot at least twice. But doing that successfully while keeping revenue growing is a challenge, especially if you’re working with little to no capital.
Giftpack is interesting not because they pivoted, but how. Chiang tweaked and re-tweaked his business until he found the right market fit and deployed some innovative strategies to get there.
Insights spoke to 28-year-old Chiang about the transformations of Giftpack—those that lay behind and ahead.
Next Article
Iterative Innovation: The Culture of Successful Startups
“We transform traditional gifting into a modern experience with AI.”
Insights: How has Giftpack evolved in the last five years?
Chiang: We started in San Francisco, where we tried to create a platform where users could send gifts online, and we’d deliver within three hours. We called our first product giftpack.io, the global automated gifting service. At the time, we were focused on creating the operations and logistics for our product, as well as expanding globally.
But something weird happened. Our customers kept asking, “Hey, Archer, what’s the best gift for my daughter?” Sometimes our customer service department was basically a gift consulting service. People really seemed to need help finding the perfect gift. Meanwhile, a lot of business customers kept knocking on my door and saying, “Hey! We want you to hand pick and personalize more than 5,000 gifts for our employees and clients.” They wanted higher feedback engagement, or regional endorsements. It can be very hard to increase these kinds of numbers.
So there was both an opportunity and an actual need. We started creating algorithms to launch our AI-based product for our corporate gifting in November 2019. Right now it’s one year—I won’t say old—it’s one year young! And we’ve already experienced a huge spike in demand. Our clients are companies like Netflix, Google, Facebook, and Cisco, and even someone new in Japan this month. These big names recognize how we transform traditional gifting into a modern experience with AI.
“That didn’t go over well for people who care about data privacy.”
Insights: Did the international expansion happen during or after your AI pivot and expansion into corporate gifting?
Archer Chiang: During. We wanted to keep exploring customer needs first—some of our US clients would say, “Hey, Archer, please create this service!” but we would still hesitate. When the Japanese clients, Hong Kongese clients, Chinese clients, and Singaporean clients all kept asking the same questions, we realized we should probably go ahead with those changes.
But gifting is also a very emotional and localized business. Our clients use gifting as different tools—sometimes for sales, and sometimes to maintain relationships. There’s a lot of reasons you might send gifts. That’s one of the reasons we need AI to make sure that we don’t make mistakes, especially in Japan. When we came to Japan, there was a lot of localization that needed to be done when we started talking to clients, especially B2B.
For example, with version 1.0 of Giftpack AI, we asked the gift-giver to give us the social media of the recipient so we could help the gift-giver analyze the personality and qualities of that person. Then we’d pick the top five gift recommendations and send them to the gift-giver via email. But we realized that we were asking the gift-giver to provide private information. That didn’t go over well for people who care about data privacy.
So we worked with Harvard’s psychology lab to create a psychology-based questionnaire. Senders and recipients can now decide what kind of data they want to provide. And they can also have a good interaction with our brand! In this way, we increased our recommendation accuracy by 40% without providing any social media accounts or keywords—though after individuals take this psychology test, they’re usually willing to give us social media data.
“If you don’t have those connections, you have to be able to hunt them down.”
Insights: How did you manage refining your AI, expanding globally, and adding a spin-off product at the same time?
Archer Chiang: I don’t have an MBA, so most of my business strategies are learned lessons. After developing Giftpack in several different countries, I realized that establishing a formal business entity from the ground up is too slow. People look at your registered capital a lot, and the procedures of cultivating a business the traditional way can cost a lot of money.
So I switched it around and decided to be very transparent. I would say, “Hey, we are an AI-driven US company. We want to do business in your country, but we need a bit of support. Our product is very unique and exciting, and we want to help you!” In other words, we give them big AI headlines in exchange for their support.
This is how we made sure we had a solid foundation in each country—we got a lot of help from local governments and accelerators. We were also really lucky and got selected into a lot of different programs that allowed us to use a local office for free or fostered local business connections.
In a certain way, everything started because of our internship program, Giftern. Every year we hire two batches of interns from all over the world. Currently we’re at thirty interns from eleven different countries. We were able to access a lot of different startup resources, sponsorship programs, and government programs via these interns.
To be taken on as either a Giftern or an employee, you have to understand the local culture. Communication, human interaction skills, and interpersonal skills are all very important. But I also look for people with deep connections—people who can get that help from big local corporations and governments. And if you don’t have those connections, you have to be able to hunt them down. That’s the most needed requirement. Finally, you need to make data-driven decisions. Our engineering team can support you by providing a lot of data, but as a manager, you need to have the ability to come up with your own ideas and draw insights from the data by yourself.
“I want to focus on getting so big, investors will chase after us.”
Insights: What’s been the biggest challenge for Giftpack so far, and how did you overcome it?
Archer Chiang: Pivoting. Our original seed business didn’t make enough money to sustain the team, so I used my own cash for almost a year and a half. I put most of my time on the product and helping with the sales process, so I didn’t actually spend much time fundraising. Raising funds is always a headache for me and so is talking to advisors, even when we have a really good track record. I finally decided that I don’t want to chase after investors. Instead, I want to focus on getting so big, investors will chase after us.
And that finally just happened! It was good timing for us. Before, cashflow advancements were a big struggle. I have a really good team that understands the situation, but it will be nice to not have to torture ourselves. [laughs] We try to make progress and impact as fast as possible to sustain ourselves. And judging by the revenues coming in month by month, I’ll be able to sustain everyone’s salaries pretty soon. So even though we don’t have investors, that’s totally fine.
But honestly, because our business is growing so fast, there’s a ton of problems popping up, especially on the operational side. A lot of different questions, different types of clients complaining, and the like, and we’re still not that big yet. We can’t cover every perspective, but we can still enjoy having these problems because it means our business growing. We’ll solve all of these complaints as soon as possible, and move forward with the next step.
Related Articles
