I have a painful confession to make. So far this month I’ve made some nasty losses on my ad spend. Ack! Ouch! Thwack!
For several weeks I’ve been aware I was deficit spending – putting more into ads than I was getting back on a daily basis. I was doing this because the last 3 months of data – which I wrote about here and here – showed the zombie box set was moving in a crazy way.
Through Dec, Jan and Feb it was selling at the rate of 1 book for every 3 clicks. Every click I paid for earned a profit of 50c to $1. KU page reads came in at 1 full book read for every 3 or 4 copies bought at 99c.
Those were amazing stats. I figured – wow. The more clicks, the more money. So I doubled down, starting this month. I ramped up Facebook ads, AMS ads, and even Bookbub ads.
The results felt pretty great – I definitely have the ability to force the book up the charts. It hit 2800 in the US store and 600 with a bestseller tag in the UK.
BUT.
Results
I was waiting for the month to end to do a proper analysis, but upon realizing the daily deficit was getting too big, I went ahead and did an analysis of the last 3 weeks, and the results are ugly.
1. Page reads disconnect
With sales of 600 books in the US in March so far, I’d expect KU page reads to be similar to December, which had about 600 sales and 400,000 page reads. Nope. It’s only been 200,000 in March. That’s half of December’s numbers. Considering I make the money on page reads – this is a major problem for profit. It works out to around 1 full KU readthrough for every 7 purchases.
In December it was 1 in 3! Twice as many. So for the clicks I’m getting, less than half the people in KU are reading now as were reading then. What the heck?
It was the same in the UK. Off 600 book sales in March so far, I’ve had only 100,000 page reads. In January, when I got 300 sales, I had 200,000 reads! so if that had been 600, it could have been 400,000. That’s a quarter as effective now. It was one full KU readthrough for every 3 sales, now it’s down to 1 in 20. Ugh. The monetary difference is big.
But that’s not all.
2. Massive conversion fail
Not only are fewer people reading through on KU, but my ads are just not converting to sales like they were. In the US in March so far, I had a whopping 4200 clicks total. That’s more than any of those stellar months going back to December.
December had 3000. Of those 3000 clicks, 800 became sales. That’s a stellar 1 sale for every 3 or 4 clicks. In March so far the 4200 clicks have yielded only 600 sales. That’s 1 sale for around every 9 clicks! Two or three times less effective.
Double ouch.
It’s not so bad in the UK, but still bad. Off 3000 clicks I had 700 sales. That’s 1 sale in about 5 clicks, compared to the earlier heyday of 1 in 3.
Combine this poor conversion rate with the weak readthrough, and I’m losing money every day. Really, what the heck is happening?
Analysis
1. Page reads disconnect
I needed more data. Previously I ran the numbers through to Dec 2019, but now I decided to go further back. I discovered some things I’d pretty much forgotten about – November was a huge month. In the US 1000 copies sold, 450,000 page reads, conversion still 1 in 4, but interestingly KU page reads were down to 1 full read through for every 6 sales.
Hmm?
I went to October. Another banner month! 1000 sales again, but page reads at 400,000, 1 full read through for every 7 sales.
Eh?
It was the same in September. 600 sales – I’m finding the point where I first ramped up ad spend – and only 220,000 page reads. That’s also 1 full read through for every 7 sales.
Exactly as it is now.
Answer – I anticipated this, but hadn’t reckoned on how long it would take to manifest. People who pick up a big book in KU may not start reading it immediately, and even if they do, it’ll take some time before everyone reading it will reach critical mass and show up en masse in my page reads.
So – the page reads I’m seeing in those stellar days of Dec, Jan, Feb, the 1 full read in 3 or 4 sales days, were likely generated earlier in the 1 in 7 days back in Sept and Oct. So it takes a month or two to really get the ball rolling. We see this as the rate increases over the months to peak in Jan at 1 in 3, then tail off afterward.
This answers the KU page read disconnect. Basically, it’s fine. It’s not the reason I’m losing money. In Sept last year, off 600 sales, I had 200,000 page reads. Right about where I am in March so far, 600 for 200,000. But in March I’m charting a loss, whereas in Sept I made a healthy profit.
So what gives?
2. Massive conversion fail
The deeper data going back to September has some bearing on this question. In Sept I had 1200 clicks and 700 sales. In Oct I had 2000 clicks and 1200 sales. Those are actually better than 1 in 2 conversion rates! That is amazing. It’s almost 2 sales for every 3 clicks. Absolutely killer!
In November I ramped up clicks and conversion dropped to 1 in 3. Still amazing, but less so – this was 3000 FB clicks and 1500 AMS clicks.
In December I cut back a little, maintained around 1 in 3, but only sold 800. I had 3000 clicks total, 50% more clicks than in October, but almost 50% less sales. Huh? Why so much weaker conversion? Had I already exhausted the market? Or were those clicks somehow less targeted, less likely to convert?
Well.
Answers
Here’s what I’m thinking now. Current weak conversion, and decreasing conversion since October 2019, could be down to a few things:
- I changed Facebook targeting to reach more people around October. Before that I was targeting primarily Kindle readers, but I broadened that slightly to include all ereaders. This might explain some of the drop-off, but it shouldn’t be so big. My audience is still predominantly Kindle readers.
- I had saturated the market – all the easy buyers had already bought. I sort of doubt this, as I see many other zombie boxsets in the market holding stronger rank than my book, with considerably more reviews – which suggests they’ve been selling a lot and consistently for a long time. The readers are there. However, it is of course possible I saturated the market for my book.
- I ran a few ads, several times, using videos that included a slightly sexy picture – these earned crazy clicks, also crazy cheap, but I don’t think they converted to sales at all. I tried that again recently, and it was the same again. Wasted clicks as people went looking for more sexy images…
- I ramped up AMS ads around October – which might have ended up including lots of bum keywords. I’m getting on top of that more with the keyword analysis tool Machete now, but it has definitely stung. I need to cut non-performing keywords rapidly, or I’m just spending money on clicks that go nowhere.
- This month I tried a new targeting approach via Facebook custom lookalike audiences. 1000 of the clicks this month came from there – there’s no way to know, but maybe they weren’t converting well because they didn’t double-target Kindle readers.
- Finally, coronavirus. I’ve read plenty of accounts in my various author FB groups that coronavirus has decimated sales. Other people are turning off ads altogether. I thought my book – as it was charting so well, and because it’s post-apocalypse – might be impervious to this, or indeed even benefit from a fascination people have currently with the end of the world. But maybe not. It also makes sense people aren’t reading as much just yet. When the panic clears a little and folks get bored at home, they may read more.
So, what does all this mean?
Action Plan
First off, I’m not going to worry about the KU pages readthrough. At 1 in 7, I’m right where I was back in September, kind of priming the pump. I believe the sales I’ve invested in now will lead to page reads over the next couple of months that should right this ship somewhat.
Ad conversion is the main area I need to focus on.
My immediate plan is to squash ad spend mightily. It would be nice to know if it’s AMS or FB delivering the lion’s share of the sales, so I may try shutting them off back and forth over the coming week. Turn off AMS tomorrow but leave FB running, then turn off FB and leave AMS running. It’ll give me some taste of what’s working.
I’m also immediately canceling all expansive targeting – so the 1% lookalike audience is gone, and I’m back to targeting Kindle readers with an image that is not sexy at all. At the same time, I’ll become more ruthless with non-performing AMS ads, and throttle them back with Machete.
Perhaps most importantly, I need to keep a close eye on conversion rates. Maybe even daily – it’s not a tough calculation to run. If they start to creep up in the next few days/weeks, it may be that coronavirus was responsible for this slowdown. If that’s the case, I’ll start to ramp up again steadily. I’ll be certain though not to do any more deficit spending. Even back in September, when I was priming the pump for future KU reads, I was also making a healthy profit because the conversion rate was so good.
I want 1 in 3 again. 1 in 2 or 1 in 4 are also acceptable. The gulf between 1 in 3 and 1 in 9 is massive. It’s good profit against unpleasant loss.
But what if conversion rates never go up again? Then I guess that is it for this boxset. It’s done a great job, but it can’t go any bigger, and will maybe tick along happily at a low spend level for some time to come. Time will tell.