So Your Sales Just Dropped By 60%, Now What?
I’ve been an evangelist for the Yahoo Store platform since 1997, and I’ve gotten a reputation for being a troubleshooter when a store is on fire. I love the idea of being the online Red Adair parachuting in to cap that oil well fire and save the day! This time of year, I get one […]
I’ve been an evangelist for the Yahoo Store platform since 1997, and I’ve gotten a reputation for being a troubleshooter when a store is on fire. I love the idea of being the online Red Adair parachuting in to cap that oil well fire and save the day!
This time of year, I get one or two retailers a week calling or emailing me, sometimes totally freaking out because their sales are way off. Usually it’s about 6-8 weeks after some event when they realize that this major drop in orders wasn’t a fluke. Now with the reality of the second Panda update setting in, my phone is ringing off the hook.
In today’s column, I’m going to give you a framework to diagnose sudden unexpected revenue drops. I’ll show you how I use analytics to quickly isolate the source of the problem. I’ll reveal several of my mistakes and examples of mistakes I’ve seen retailers and developers make and how we fixed them.
First & Foremost, Don’t Panic!
It may be easier said than done, but first, don’t panic. Keep a calm head and isolate the problem. Half of the time if you can find the sales leak, you can patch it pretty fast. Once you know what’s off and where it is, you have a better idea what to work on.
Is it one big thing or death by 1000 cuts? Is something broken on your site or is it outside your store? Is it changes in your marketing channels affecting traffic, or industry trends, or the economy?
Usually a drop in sales is a combination of lots of things. It’s rarely just one little thing wreaking havoc, but retailers don’t notice until they feel like sales are half what they should be or sales have dropped off. Sometimes, the event that killed your sales is something that’s been happening for a while, and the update just pushed sales into the danger zone where they got noticed.
Place A Real Test Order Right Now!
The first thing I do is place a real test order with a real credit card and see if something is obviously wrong with the checkout process.
When you’re troubleshooting, make sure you’re using the same browser as most of your customers (Internet Explorer) and use the operating system, version, and even screen resolutions that most of your customers use. I love Macs, and I’m running Chrome on my Mac right now. The major problem with that is you can’t see certain error messages that Windows users running IE see.
For example, on one cart redesign I did recently, one logo in the secure checkout used the path to the normal version of the logo (not the one on the secure server) so some browsers gave error messages warning that some items were not secure.
You could only see this on the published version of the cart and only in some browsers, but we caught it about 15 minutes after deploying around 11pm one night. That error alone could totally nuke your conversions for the next day!
Houston, We Have A Problem…
Usually, when there’s a problem in the cart/checkout, especially if you have a high order volume and/or lots of repeat customers, someone will give you a heads up that your cart is down. Make sure your customer service folks know to let you know immediately if there’s a problem.
Smaller stores have to be more careful. If you get less than 40-50 orders a day, you realistically don’t have enough order volume to get a phone call or an email as soon as your cart goes down. If your credit card processor goes offline or has a huge lag, not every prospect is going to let you know there’s a problem. Most folks just move along to the next guy selling exactly what you sell.
For short-term trouble-shooting, Yahoo! Web Anaytics runs in real time, so it’s easy for me to see what the immediate problem is. Is it traffic? Or folks are adding to cart but not checking out. If I have a disproportionate amount of dropped carts, it’s usually something in your back end. For longer term issues, we have to dig in a bit.
Look At The Numbers Over The Long Haul
I’ve learned that many times you can’t trust a retailer to tell you what’s really going on. When someone cuts off the money machine, retailers start freaking out and lose all objectivity. There are bills to pay, cash flow is tight, and orders are way off. Remain calm and we’ll get through this together.
After making sure the site works, I build a simple model of how the business works by the numbers. I dive into my stats in Yahoo! Web Analytics and/or Google Analytics and look at the previous years to see the normal cycle of traffic and sales over the selling cycle of a full year. Looking at multiple years will show you patterns.
For example, how many orders should Tom get this July, if July is traditionally about the 75% revenue of normal April traffic? If he maintains his 2% conversion rate with a 50% completed cart rate with an average sale of $100, how much traffic does he need to get these orders?
Examine Long Term Trends Over 90 Days To See The Big Picture
I’m not a math-person at all, but I like to look at rolling averages over time to smooth out the bumps that happen over time. I also like to compare periods of 90 days (at a minimum) when I’m trying to see what’s happening.
Huge variation in sales in normal comparing day-to-day. Holidays fall on different days every year. Huge news events can distract folks. Freakish weather can disable large parts of the country for days at a time. People get paid on different dates than last year. Take all of this into account when “sales are way off!”
Many times retailers look at their sales in the first half of the year coming off a big fat Christmas season and exclaim, “Our sales are down! Our sales are down!” Well, your sales are down from your peak, but how are your sales are compared to the same exact month last year?
The default date range on Google Analytics is the last 30 days. And if you click compare to past, it compares that range with the previous 30 days which drives me nuts. I guess in a dot-com tech start-up universe, you might need to beat last month’s numbers every month, but retail is a little less insane.
Traffic Metrics I Watch
For traffic metrics, I compare visitors and revenue from a specific channel over past 365 days to the previous 365 days. If there’s a big drop I dig in deeper. I look at individual keywords and/or entry pages to see if something is broken that I can easily fix.
Conversion Metrics I Watch
I look at the following over time for seasonality and compare year-to-year:
- Unique visitors
- Product view rates
- Add-to-cart rates
- Proceed-to-checkout rates
- Completed cart rates
- Number of orders
- Average order size
Remember: Correlation Does Not Imply Causation
Just because you changed something right before sales tanked does not mean that is the cause of your problem. It might be worth looking into, but remember that correlation does not imply causation.
If It Ain’t Broke, Don’t Fix It
Many times a problem pops up because somebody changed something and didn’t realize what the unintended consequences would be. Maybe it’s a store redesign gone bad which affects traffic and/or conversion. More often, it’s checkout-related.
You need to know what’s different today compared to when sales were good.
Compile A List Of All Site Changes
I like do the following:
- Keep a log of changes to your site.
- Archive HTML copies of VIP pages
- Take screenshots before you make major changes.
- Mark big changes in your analytics package for better reports.
Changes That Could Affect SEO Traffic
- Have you radically changed your ROS NAV and/or internal linking? aka, lose 20K links…
- Have you increased the size of your boilerplate template?
- have you added NOINDEX, follow on all your low content pages
- Have you doubled the size of your site with a datafeed?
- Are you sharing your content via shopping feeds?
Here’s an example of a traffic drop caused by a site redesign:
A retailer switched from a run-of-site template with 20 category links to a flyout menu navigation with a links to over 250 different pages.
On a site with 20,000 pages that radically redistributed the link popularity, but it also quadrupled the size of the boilerplate of the template. My take was that Google saw so many really empty category pages and very thin product pages with more template than content and removed most of his site from the index.
You want to be really careful when you’re doing a site redesign that you don’t have so much boilerplate that it overrides your really thin category and product pages.
How Close To The Ground Are You?
Occasionally, I’ll see someone who doesn’t have so much traffic that they’ll drop from 500 visitors a day to 250 visitors a day. And that’s one of those situations where it’s so little traffic that the slightest change in algorithm or page layout on the Google search engine results page can kill you.
Changes That Could Affect Your Add-to-Carts
Have you changed your shipping rates or “free shipping” model?
Another thing I look at is whether you have changed your shipping model recently. Did you used to have free shipping and now you don’t? Have you increased your shipping rates?
Have you changed prices on best-sellers?
One example I use pretty often is a manufacturer raised our cost on one SKU, so we passed them on to the consumer and sales of that SKU instantly dropped by 60% even though all our competitors did likewise. We actually made more money, so you need to watch revenues and profits. Like my daddy used to say,”It ain’t what you gross, it’s what you net.”
Sometimes it’s more than price…
In one of our stores, all our competitors were offering a free gift with the purchase of a MAP (minimum-advertised price) protected product. It was a pretty lame free gift, but when we added that, sales jumped up. Are any competitors outdoing you here?
Changes That Could Affect Your Checkout Completion
Have you done something in your cart that would have possibly impacted your sales?
For example, one store I helped last year had this “Dropped Cart Catcher” that would offer you a 5% coupon when you were bailing from the site, but when you elected to stick around, if you were in the cart, it would “forget” to give you your discount. Another example of something implemented but not tested by the developer or the retailer.
Here are some other quick additions I’ve seen kill checkouts:
- Added COUPON fields
- Requiring a checkbox for terms and conditions
- A/B testing software that was slow to render the pages
- Increased shipping rates
- Handling charge or other extras to monetize shipping
- Shipping calculators on sites with a free shipping model
Adding All These Third-Party Features Sounds Cool, But Is It?
Try not to run too many different external services in your cart. It’s just one more thing that can mess up and prevent your customers from giving you money!
On one of our stores, we were in the stone age and processed credit cards offline just up until a year ago so we could catch the folks with declined cards or when there was a problem with the gateway.
Also, I don’t use the UPS shipping calculator, real-time inventory, external address correction software or many other features that require another company’s server to work for my customers to give me money.
Your Homework Is To Place A Test Order
Right now, I want you to order something from your online store.
- Go to products.google.com
- search for site:yourdomain.com to show only your products
- sort by price from lowest to highest
- find the second cheapest item you sell.
Stop. Don’t click that link!
Now, try to find this item on your store by browsing or searching and then order like a customer would with a real credit card. Start making a list of things that might be hurting your sales!
Keep Your Head While Others Are Losing Theirs
It’s a little too late for folks already on fire to do this, but here are some tips that may help you avoid the next sales drop, or handle it better and sooner than you would have otherwise.
- Have your finger on the pulse of your store’s metrics.
- Place a real live test order once a week.
- Know which channels drive traffic and more importantly, revenue and profit.
- Make changes slowly and one at a time so you can see the impact of each
- Network with platform providers and fellow merchants who share the same service.
Don’t take double or triple digit growth for granted. If your business model is dependent on growth like this, you’re gonna have some problems! Sock some cash back when things are fat.
Don’t panic and you’ll make it! I’ve been there. It sucks, but you’re tough enough to handle this!
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.