Giving users back their data is good for business
There’s this prevailing idea that in order to sell, you need to keep your big people data and algorithm sauce completely secret. It’s a model that works well for passive interactions like ad targeting, and problematic when your personal data is used by major corporations like Facebook and Google. To quote: ‘If you’re not paying for the product, you are the product.’
Except this method doesn’t work with conversation-as-commerce. The moment you’re interacting directly with a user, the curtain of invisibility — fundamental to platform data-tracking, is lifted. Users are constantly reminded of your presence and trust plays into every interaction.
We found this out with Kip, our AI shopper penguin on Slack. Through testing, we discovered people will shop more if they know exactly how you come up with product matches, and are more forgiving when you get it wrong.
Most importantly, people don’t make purchase decisions unless they know they’re getting a good deal. Unless you tell a user what your product is worth, no customer will buy because they can’t judge the value-for-money. This applies to everything from restaurants to fashion to luxury hotels.
By making your algorithm opaque, the user doesn’t know the value of your service. They think someone, some company, is cheating them. Even if you have the fanciest AI on the planet, doesn’t matter. The sale never happens, and your conversion rate drops.
Just to say you’re transparent isn’t enough. You need to spell it out. You need to say: Hi we’re Kip and here’s how we sort your requests:
1. Your purchase history
2. Price points at 3 different aggregates: 40%, 60%, 80%
3. Product reviews/ratings
4. What other people who searched for similar things ended up buying
5. Lucky wildcard
Users should be able to opt-out anytime, and companies should actively encourage user participation by asking them what they’re looking for. Instead of building filter bubbles via guesswork, ask your users what they like and dislike, and let them edit their interests.
We thought we were pretty good on this, especially after being granted a public benefit corp status. Then someone outside of startupland pointed out: We get affiliate revenue from each item sold via our chatbot, how trustworthy are our results? Would we show preference over one retailer against another who gave us a better revenue rate?
That gave us a pause. Even if we said we were serving up the links based on user’s best interest, users could disbelieve us. It was an impossible case to prove. In the end, we can only say this:
Being transparent helps create trust. Trust is an on-going process between our services and our users. In exchange for trust, Kip gives the best possible experience. We’ll rather have long-term recurring sales as a trustworthy bot that has your best interest, than short-term high revenue by offering bad deals and destroying user trust.
Once the trust is gone, it spreads like a virus and becomes extremely difficult to repair. Therefore, it makes good business sense to be as transparent as possible with user’s best interest in mind.
TL;DR: Be transparent. It’s good for your users, it’s good for business.