Peaks and Valleys in Consumer Sales

I'm curious what kindof delta people see

[isellstuff]

I've been running an large e-commerce website for over ten years that covers a large selection of consumer products for all buying seasons. One thing that constantly frustrates me is the peaks and valleys in the daily merchandise sales. So for instance, one day I might sell $30K in merchandise and the next day I might sell $15K. Sometimes this delta is caused by a reduction in average order size, sometimes it is caused by lower conversion rates.

So take this week... I saw a 50% - 60% delta between total merchandise sales on good and bad days.

I'm curious what everyone else experiences? Is my problem that the number of daily sales needs to be in the 2K, 5K, or 10K unit range to see less of a delta between good and bad days or is that just the way e-commerce rolls? Obviously some days people just don't feel like spending money, but I find this drastic delta to be very frustrating. It even affects my moods. I'll have a couple great days and think I"m going to rule the world, only to have a string of really bad days. In the end, they generally even out to "just mediocre sales" for the week.

[fathom]

I'm more of a "didn't need to declare bankruptcy today, type of guy" so the impact of hourly, daily, weekly, monthly or yearly trends don't phase me much.

I get to work while on permanent vacation so unless revenue dries up completely for months my mood is great.

[tangor]

Might look at days/trends in the concept of when people get paid, holidays, or other criteria. First/Mid/End of month usually shows spikes for me (when people get paid).

[RhinoFish]

Segment past data by day of week, so that roughly half the time you can dismiss highs and lows as following a normal weekly pattern.
Take some of the randomness out, will make you realize the up and downs are mostly natural.

Then look at weekly patterns, and how smooth they are, even with the known ups and downs, and you'll come to see the ups and downs cancel each other out.
So any one of them by themselves,up or down, is neither a time to celebrate or cry.

Data and patterns take the emotional out of the roller coaster for me.

As an agency, when a client asks us why the day was slow, we look across all clients, often finding a pattern there too - can be relieving to the client (and us) to know that things were generally slow (or fast) for many that day.

[isellstuff]

As an agency, when a client asks us why the day was slow, we look across all clients, often finding a pattern there too - can be relieving to the client (and us) to know that things were generally slow (or fast) for many that day.

Wow, that's awesome data to have. I spend a lot of time in a spreadsheet that charts daily revenue for the last 5 years. So I'll sum revenue over the last 7, 14, 28 days, take averages, compare with past years, etc. 2015 trending has deviated from 2014. Actually, I've found sales to not trend in the same way, YOY from 2013-2015. Prior to this period, it seemed like sales were a lot more predictable.

[ecommerceprofit]

I would like to say I don't worry much but I do too...even after 20 years of e-commerce sales. I'm getting better at looking at the big picture - sometimes we get crazy busy days and sometimes super low days...good advice to look at weekly/monthly data rather than daily data. Always seems like when I invest in expansion / new employees sales go way down for a day or two and then recover...

[RhinoFish]

The 21-day vs 7-day moving average of revenue is by far my favorite metric for trend analysis.

Regarding the 21 days...
14 isn't enough to set a concrete baseline...
28 is so long that it slows response time.

7 and 21 are both whole week multiples, making the MACD of them, powerful for me.
And my sanity.
:-)

[tangor]

Depending on how "immediate" your desired trend view, the 21/7mavg comes nearest to enough data for some consideration.

My report (custom) does that as well as 12 other looks at data/trend, mostly using historical data as well as most recent. And can do nice charts, too. :)

But what I have learned since the middle of 2014 is that everything is generally "down" to previous years. Not terrible, but not "up".

Edited to add: This speaks only my general results over a number of sites, many of which are mfg/unique and thus not truly your usual ecommerce sites. Keep that in mind.

[isellstuff]

But what I have learned since the middle of 2014 is that everything is generally "down" to previous years. Not terrible, but not "up".

This is my experience as well. Exactly my experience! I saw an RKG video blog post that said the exact same thing. Actually, I think they made some sortof comment about an inverse relationship with Google's stock price and their own sales trends.

Whenever I talk to Google Rep's, I always throw that "mid 2014" thing in their face. e.g. You know, I was making you all a lot more money before mid-2014. You know, I felt like helping you guys right up till mid-2014, now I have no incentive what-so-ever. I encourage everyone to do the same. They screwed us e-commerce people by redirecting most high converting traffic through Google Shopping widgets in order to keep Google Shopping advertisers happy enough such that they can continue to fleece them with a mixture of un-targeted traffic and targeted traffic. They pushed the natural SERPS below the fold, on purpose, IMHO...

And just so we are clear... You can get early stage buyers on products with a long buying cycle. But when they get ready for the purchase, if they start their search on Google and don't mention your domain, then you won't see that traffic again because it is being routed through Google Shopping. This is what happens when you give Google your conversion data. They can predict user behavior, very, very well.

[Essex_boy]

They pushed the natural SERPS below the fold, on purpose, IMHO

[Essex_boy]

I do agree with the comment