135 Greg Mercer on the Best Products to Sell On Amazon – Part 1 of 5
We have Greg Mercer on the show again. You can listen to our previous interview on product research, as well as one on supplier negotiations. Greg studied civil engineering at university and had a corporate job that he hated. He began selling on Amazon as a break from his day job. He managed to quit his day job and just do FBA full-time. He did that for about two years when he was frustrated by trying to find products to add. The best way to scale your Amazon business is by adding more products. Greg didn’t have a lot of capital to throw around so he wanted to find ones. Out of this need, Jungle Scout was born. Now he joins us to help us find the best products to sell on Amazon.
Today, Greg is still selling on Amazon. He has released a few products in the last few months. He’s been working on Jungle Scout, and that has expanded into a quite a tool for Amazon sellers. There is Jungle Scout, which a research tool. Jump Send is a deal site to get you additional sales, as well as a follow-up sequence. Splitly is an AB testing tool for Amazon sellers. Fetcher, which is profit analytics. It calculates what you’re really making after refunds, promos, etc. All the numbers Amazon likes to hide from you.
Let’s start with the one thing that every struggles with: finding something to sell. What is your first step to finding the best products to sell on Amazon?
That’s a common issue. Everyone knows how good of an opportunity Amazon is, but it’s finding products to sell that is a struggle. The best products to sell on Amazon are ones that have existing demand, that means Amazon customers are already searching for it. You want products that have low competition and that have good margins. Those are the main things. Other things you may want to consider are whether they may infringe on any patents, and they don’t need to be licensed. Think of liability; if a person can hurt themselves with it, you may want to steer clear. Lighter, smaller items are generally less complicated. They are easy to ship and you don’t have to worry about oversize storage limits.
Jungle Scout was created to solve that issue, but you can look on the Amazon’s best sellers page. You can get ideas from Pinterest, look at what people pin a lot. You can hang out in big cities where trends start first. Once you do that, make a list of product ideas and go to Amazon. There is actually a free way to find out how well a product sells. You can click on a listing, then look at the best sellers rank under the product description. Then you can go to junglescout.com/estimator. It’s a totally free tool, you don’t even have to put in your email. You put in that sales rank and it will give you an estimated amount of units that product sells on a monthly basis and see what the demand is.
You say you want high demand and low competition, can you define that in some way, and how do you find that out?
For demand, you want to look for products that are already selling on Amazon. A beginner mistake is that people “know” that a product will do well if it gets on Amazon. A small percentage of the time, that might be true, but more often than not people are wrong. It’s much safer and less risky to go with something that is already selling.
I want to see 2000 units a month, being sold on Amazon. Let’s use a coffee cup as an example. If you search “coffee cup” on Amazon. Then take the top 10 listings, or however many are relevant. Let’s say 8 are selling coffee cups. Then click on each of the listings, get the best sellers rank. This is helpful because it tell us how well this product is selling. This number, by itself, is very difficult to interpret. However, at Jungle Scout, they have come up with an algorithm that can estimate how many units are sold based on that number. It changes on a daily basis and they have a full-time data scientist that is always updating this. So, get that number for each listing, find the units sold on Jungle Scout and add them up. If it’s more than about 2000, then the demand is there.
One issue that many people have had, is regarding the monthly sales number. Some sellers have noted that the estimate Jungle Scout shows differs than the actual numbers when they lookup their own products.
The first thing is to look out how they drive these algorithms to estimate the sales. Depending on the category, they collect between 200,000 and 500,000 data points every month for that category. This is the relationship between the unit sales and the ranking number for that day. Then they run a regression analysis and they come up with a line of best fit to estimate the sales based on the rank.
The best sellers rank changes on an hourly basis. The way they estimate sales is that if a product continues to sell as well or as poorly as it has for the past few days, this is how many units will sell in a month. If, last week, your product was selling 10 units a day, but this week is selling 1 unit a day, Jungle Scout will estimate based on the 1 unit per day. So you’re sales might be 60 units that month, but Jungle Scout will only estimate 30. It’s the best they can do with the limited data Amazon gives out.
Some people will get on there and see their products are 10% more than Jungle Scout’s estimate and will conclude that you need to add 10%. That’s not true. If you look at the regression analysis, there are some points that run above the line, and some below. They’re taking the average of hundreds of thousands of products in a particular category. So, your 1 product may not fall on that line, but if you average the whole category, it will be on that line.
People seem to put too much faith in tools. You can give me the best painting supplies in the world and I still could paint a good picture.
Exactly. People get caught up, too much, in the tools. Keep in mind this is still just an estimate. You’re using this tool to determine a ballpark range on a product’s sales. Jungle Scout may estimate that a product does 900 units a month. In reality, it might be 800 a month, or 1000 a month, but you know it’s in that range. It helps with forecast and it help determine if there is good demand in there.
Let’s go back to the 2000 units a month number. I had a product that, on Cyber Monday, sold 103 units. Now, the same product is selling 4 or 5 units a day. How do you account for that, especially this time of year when sales volume tend to be low?
This is difficult. One tool that helps is Google Trends. This tool allow you to see how a search term has trended over the years and seasons. This is a fairly good gauge of how items will sell on Amazon. As many people know, Greg has done public case study selling bamboo marshmallow sticks called Jungle Sticks. Based on Google Trends, you can see how the sales have changed based on the seasons. January to February are the slowest times. July and August were the highest times. And if you look at the sales, you can see that matches up. So can look on Google Trends to determine if this is a high season or a low.
The reason I like to use the 2000 or 3000 units, is because people like to answer “It depends”. It’s too arbitrary if you’re a beginner. At the end of the day you’re looking for the item with the biggest spread between demand and competition.
As you know, sales tend to spike in December, plummet in January, then even out the rest of the year. Would you try to take account of that?
If I was a complete beginner looking to sell my first product on Amazon, I wouldn’t worry about that. That’s more higher level strategy. Focus on getting your first product up on Amazon and learn the rest later.
If you’re already have your products on Amazon, and you’re trying to figure out forecasting, that is a good idea. Two good resources are Google Trends, and Keepa. Keepa has a really nice, free database of how sales rank has trended. A lot of products have two years or so of data. You can look at the and see how the sales rank has trended over the months and seasons. You can try to start estimating how well your product is going to sell.
Some products you can tell by common sense. If you’re selling lawn products, then the summer months are going to be the best. Other products, like the marshmallow sticks, it’s not as clear when they’ll sell well and Google Trends can help with that. If Google Trends shows there is twice as much searching for marshmallow sticks in the summer months, then you know to order a little extra inventory.
How do you measure competition?
Reviews are a great indicator of competition. That’s probably the biggest thing to look at. On top of that, the quality of your competitors listings. If they have a poor listing, like one picture, a really crappy title, than that is someone that would be much easier to outrank. As opposed to someone with a really good listing.
The first thing to look for is how many reviews they have. Older, more mature products that have been selling consistently well, are harder to outrank.One way to tell how mature a product is, is how many reviews it has. An older product that sells well, is going to have more reviews. A product with 1000 reviews is going to be much harder to outrank than one with 15. A rule of thumb is to look for something, where 3 or 4 of the top reviews have under 50 reviews. That signifies that it’s probably a young niche.
Tell me about the relationship between the average review and the number of reviews. I had a product that had a 4.9 average but only 22 reviews. It was selling quite well against competition which had 700 reviews. Is that a one-off thing or is there a correlation between the average review?
One thing to understand is how Amazon ranks the listing. They use keyword relevance. The sales velocity probably makes up about 50% of the algorithm. That would be the number of sales per day. Another factor is the conversion rate of your product. Now the sales velocity and the conversion rate depends on a number of factors. Those including the quality of your pictures, the price, the social proof, the average star rating. If you competing against other listings that have a lower rating, then you’re probably have better conversions and more social proof. People would much rather buy a product with a higher average rating than one with more reviews.
I’ve noticed that if a product goes from 4.9 to a 4.6 average, the conversion almost halves.
Yeah. Visually, if you have a 4.9 average,Amazon displays 5 stars. But a 4.7, they show 4 and a half stars.