How Groupon is Losing $1,117,808 Every Single Day | Hacker News Reader

How Groupon is Losing $1,117,808 Every Single Day

blog.priceintelligently.com

25 pointsaaronwhite16 comments

The title here seems a bit sensationalist. I was expecting to read a story about how Groupon is bleeding a million per day. Instead, this is an article about how one company's analysis has concluded that they could be making $1,117,808 more per day.

A missed profit is not a loss. It's a missed opportunity, sure, but I don't think calling it a loss is appropriate.

Drake - you're right it is a bit misleading, because I would equate a loss as more from marketing, waste, or mismanagement. Yet, I'd also equate missing out an opportunity as a loss as well, albeit not in the traditional sense.

Totally agree, I was expecting the usual bash Groupon and daily deals article.

Cant wait to see where they are in 2 or 3 years

I agree that "A missed profit is not a loss."

For example, Amazon is in my opinion a company that doesn't focus on increasing profits, especially in product areas where they are trying to push out competitors. Groupon could be artificially holding their price low to force out competitors.

This seems like a spammy advertorial to me. There is nothing in the article to back up any of their sensationalist claims.

As far as I can tell, they ask customers how much they would pay, and draw a pretty graph of that. What evidence do they have that people's responses to price questions are an accurate indicator of what they would actually spend? What have they done to address selection bias?

Great points, rtb. We've actually done a lot of research on validating our process and algorithm. I addressed this in another comment, but we've run this process with commodities (very marke driven products) to ensure the process is sound (check out more info here: http://blog.priceintelligently.com/blog/bid/153543/Pricing-i...).

We've also started working on honing more and more validity checking into the model to filter out as much noise, outliers, and selection bias as possible.

The next step for this particular angle would be to test the deals out at the higher price point. Past customers of ours have definitely seen changes. Appreciate the feedback though. For the next part of the series we'll make sure we outline more about the foundation of the methodology and the validity behind it. Happy to hop on a call to outline, as well.

Thanks for the thoughtful reply. Your gold example doesn't really prove the approach sound though. If you ask a gold buyer how much they would pay for gold, I would expect them to be able to give you an accurate answer immediately.

However, if you ask a consumer how much they would pay for a discretionary purchase, I would not expect their answer to match reality.

Let's say you asked me how much I would pay for a BMW. I would probably never buy a BMW at any price (except if I thought I could sell it on immediately for more), but I would happily name a price if I was asked. This would distort your figures.

I just don't believe that what people say about prices matches what they will do (except for traders in commodity exchanges, which I think is a very unrepresentative example v.s. consumers visiting a restaurant).

(Thanks for the offer of a call. Luckily, I don't have the stress of setting any customer facing prices at the moment, so I'm not a potential customer, but thanks anyway.)

Surely a fairly core issue is the type of customer who buys from Groupon.

If Groupon deals are primarily purchased by a fairly small core of regulars (in my very limited experience a few people I know use groupon a lot, most never do), then increasing the price one day might just put them off buying tomorrows item. Every person has a finite amount of disposable income.

Exactly. There are so many other factors regarding pricing: Long Term Value, and Cost to Acquire being one of them. It is common to break your audience into 'bands', and learn the trigger points for each band: cost to acquire a customer in this segment and how likely they are to make repeat purchases, at which price.

Simply put, the majority of your audience may make repeat purchases, but only if each purchase is a separate price. This is much greater value than a one-off purchase.

Apart from the other issues people have listed here, I'd take issue with their 50/50 assumption. We've been offered 70/30 right off the bat without any negotiations by Groupon and its competitors and that was as far back as late 2011.

In the qualitative research we found all kinds of different splits, but the consistent one was 50/50, with a lot of articles mentioning that the 50/50 was actually worse for the merchant, because Groupon then included processing fees.

At the top level though, we're more focused on the price sensitivity information, so even if we change up the split, we're still talking about a good amount of revenue being left on the table.

Interesting article, when I read the headline my thoughts were: What about Living Social the company that Amazon has written off as a dead investment that cant IPO and will be gone well before Groupon?

There is a lot Groupon can do to perform and the huge war chest they built from the IPO has given them a great runway to do so.

Living Social probably runs into the same exact problems, but we'd have to test Living Social users specifically. What's interesting is that in the qualitative research, we found merchants speaking favorably about Living Social vs. about Groupon. Customers seemed to be a bit more affluent, as well, almost indicating that a lot of "BuyWithMe" users switched over to LS, instead of focusing on Gilt deals.

I think this shows I've read too much about daily deals.

are you sure people would have actually bought at the higher prices? The Groupon revenue is from actual sales, your data is from asking people. People often say they pay a lot more than they actually will press the "buy" button for.

Great question. So, we've definitely done a lot of validity checking in the form of pricing out commodities (gold, beans, etc.), but the customers we've worked with have also seen some great lift from the changes guided by the data. You'll never actually know until you put the price out there, but the foundation is pointing towards yes.

That might be our next form of validity checking, pricing out a daily deal, and then selling it. Splash pages aren't hard to build out, after all. More on how we priced out gold here: http://blog.priceintelligently.com/blog/bid/153543/Pricing-i...

Interesting analysis; However, how does one factor in sensitivity in this analysis? Wouldn't one want to know how the demand will be affected if the price is increased?

Reply on news.ycombinator.com