My texture generator sales are driving me nuts.

It's worse than watching the stock market!

Throughout 2007 I made one sale every 2 days, consistently.

The end of december I made a change to my campaign. I started running a second old campaign which had been doing the same sales as the one I was currently running. I hoped it would double my sales. I also reduced what I was paying for impressions with the original campaign and it seemed like doing so reduced my impressions a little but decreased my spending a lot.

My sales suddenly shot up to 1 a day, sometimes 2. More often 2 as the end of the month approached.

Then as suddenly as the sales jumped, they plumetted. Suddenly the texture generator sales stopped completely.

I'm still not sure what caused this, but it may be partly due to the fact that while I thought my impressions remained constant throughout january, they actually were dropping like a rock due to my change in how much I was willing to pay. By the time feb rolled around they were near zero, for one of the two campaigns. The other however constinued to run normally.

Since then I've been in damage control, talking to the owner of one of the sites I run my ads on, trying to figure out what happened with my ad impressions there going up and down, and putting my bid back where it was, trying to get back those lost sales and get my advertising budget back to where I was only spending 50% of my profits instead of all of them.

I think I've done some good optimizing of my campaigns now, so I expected to start making a profit again here soon instead of losing money. I made a bit less than sale a day over the past few days, and I've raised the price of my program a bit so I can advertise a little more.

But then today happened.

Over the past few hours I've sold SIX copies! I've never sold six copies in one day before. And I don't know why it's happening. I emailed the people who bought it today and asked them where they saw my ads, and two so far have replied. One said he just searched for textures in google and found my site. The other said he saw it months ago and just got around to deciding to buy it.

I HOPE that some more people reply and tell me that they actually saw my ads on cgtextures.com where I've recently increased my advertising budget after reducing it and doing other adjustments to try to figure out how many of my sales come from there and optimize my costs. Cause if they don't tell me that then I don't know what to think. Maybe it just takes 2-4 weeks for people to actually buy the product after seeing my ads. But that wouldn't explain the sales spike in January. And it wouldn't explain the sudden drop in February either. A gradual drop maybe, but not a sudden one.

And what I suspect will happen is tomorrow I'll start getting one sale a day again. Which is good.

But... So frustrating!

Just checked out http://cgtextures.com/ (nice link BTW) and an ad for Seamless Texture Generator came up on the second page I visited.

As you no doubt know, wild fluctuation are common in statistical systems with a small sample size, so it's not surprising that you get those sort of results. There may also be seasonal factors which can affect results in various ways (holiday periods, etc).

EDIT:
As GW mentioned below, it's interesting hearing about your experiences in selling your product, particularly in regards to pay-per-action advertising. I have a world editor nearing completion and am starting to think about my options for selling it.

I cant comment on the content of your post, I have no experience with online ads. But I'm glad your posting stuff like this. I'll soon be doing this myself and find it very interesting.

Surely there's a good degree or randomness to this anyway. You could sell 20 in one day, and not sell another one for 20 days. There's too many factors to predict sales of something like this swifty.

Nice info though!

But it's not random. Like I said, sales were steady at 1 every 2 days throughtout 2007. Of course there were fluctuations. But they weren't very big ones. I'm not sure I ever went four days without a sale. If I did it was rare.

Also I think I mentioned in an earlier post that I calculated that I needed about 50 downloads to make one sale. Since I based that on the course of a year that's a pertty accurate number. I also calculated that for specifically targetted sites like this texture site, the sales ratio was more like 7:1. So much better. These are averages over long periods of time, so while it's possible they fluctuate a bit, I'm sure they don't fluctuate that much since the steady sales indicate they don't.

Makes me wish I developed the big idea further!.

How much profit you made on this little venture all together then?

Its exactly the same for me, one week selling really well then the next not a sausage.

Jan & Feb seem to always be the best months, mid-summer being the worst (makes sense but doesnt explain the weekly phenomenon).

Hi there
Just a thought, but can't you have some sort of unique link or reference or whatever for each place you advertise so that you know straight away where purchasers are seeing your ads? That way if most of your customers are seeing one ad and only a few another you can adjust your budget accordingly.
Regards

It hasn't been that profitable. I have to put 50% of my profits back into advertising. So if I sell 15 copies a month, and make $22 profit on each, I make $4K a year off it, but then have to turn around and spend $2K of that on advertising.

I started selling it in 2005, so I've only made maybe $6-8K off it so far. But it continues to sell steadily, so I can expect it to continue to sell for at least 2-3 more years. And hopefully my changes in advertising will increase my gross to maybe $10K a year so I'll make $5K in profits.

I'll need to update it eventually though if I want to keep selling it because it doesn't work with 64bit operating sysems.

BigH:
Not really. I mean technically I could have a totally different website for each place I advertise, but that's not really realistic. The ad links go to my website so they can see the demo and list of features, not sraight to a purchase page.

You can't expect sales to be that consistent, surely?

Sales of everything fluctuate.

There are factors like a lot of people dont get paid until the end of the month, etc., too.

I have not noticed any trends regarding more sales near the start or end of a month. But I have not analyzed the sales to see for certain if that is the case.

And yes, the sales fluctuate. But... well, look at the chart for yourself:



These are the sales of my product by month, since I started selling it March 15th of 2005.

What can you tell from this?

Well, sales of clearly picked up over time.

And you can see that sales are fluctuating less as time goes on.

And that since the start of 2006, only once have I sold less than 8 copies a month. And that in 2007 I sold at least 12 copies every month except one where I sold only 11.

You may also note that my sales in Feb aren't that bad. That's because I sold no copies the first week and then went nuts adjusting my advertiisng trying to find the problem, because I've never gone a week selling 0 copies. I also experimented with different budgets, and added new banner shapes. As a result I still sold a decent number of copies that month in the end, but I also spent more than I made on the software, so my profits for that month were effectively zero.

And don't be fooled by the low bar on the end for March. We're not halfway through the month yet, that's why that is where it is.

Now, you may be wondering about those spikes in sales. Are those around the same time? The answer is, no, those aren't even close to the same months, and it is not some annual fluctuation.

The first spike, in April 2005, was due to the software just being released. It was bought by several Blitz users, it was posted on Garagegames, I think I even posted a link on CGTalk.com.

The second spike in August 2006 was due to CGTextures.com opening up. They too posted on CGTalk and my ads ran there automatically due to a keyword targetted campaign I was running.

The spikes in May 2007 and Jan 2008, I don't know the origin of. I don't even rmemeber having a spike in May. I guess I'll try to use google and search on my product's name in those time periods if it allows me to do that, and see if I can find a mention of it somewhere. I may also have made a chnage to my advertising then which resulted in more sales but lower profits. I'd have to combine this chart with my google adswords statistics to determine my profit for each month.


If you want to see the sales for a specific year month by month just change the filename of this image. This is 2005:



Note that the individual charts for the years are not to the same scale.


Anyway, so as you can see, the sales do fluctuate, but not that much. Between about 12 and 16 copies a month right now. That's +-12% of the average.

Oh and one more thing. I just did some new calculations on my advertiisng with some new numbers I realised I could use, and these new numbers give a different story for how many downloads I need for my pay per action campaign where I pay 25 cents per download.

My previous calculations showed that I was getting 1 sale for every 40 downloads on that campaign. At my previous price of $24.95, with profits of about $21 a copy, I could spend about $10 on avertising per sale, so 10/40 = 0.25.

But I calcualted that based on my total sales during that time period. I just remembered though that I also tracked the sales that resulted from those same clicks during that period, paying $2 for each of those. But when I divide the tracked sales from those impressions, which are less than the actual number of sales I did, I come to a quite different conclusion, and that is that I have to have around 200 downloads to make one sale.

The reason for this descrepancy is likely because my pay-per-download campaign doesn't target the same sort of demographic as my other advertising campaigns. I might make one sale for every 5 downloads from users of cgtextures.com, but some site called photoshoptutorials would have users which might not need a texture generator as much and are just trying it out of curiosity.

Whatever the reason, at my current price of $34.95 for the texture generator I make $30 in profit on each one, so I can spend $15 on avertising, but if I divide that by 200, then I come to a budget of only 7.5 cents per download.

The true number of sales I make per download for that ad campaign is likely somewhere between those numbers since I doubt the sales tracking system is perfect, but I suspect it's closer to 200 than to 40. Which means that I've been spending a lot of money on sales which may only have been breaking even, and at worst have been eating into my legitimate sales.

But I'm not entirely convinced of that yet. I'll have to do more adjusting of my avertising to determine the real effect those impressions have on my sales.

One test I'm doing right now is instead of paying per download, I'm paying $15 per sale. I was getting 500,000 impressions a day with the pay per download. With this new setup, I'm only getting 60,000. And that number may drop. I don't know how google assigns impressions for the pay per ads, but I do know they don't neccessarily adjust you down immediately if you start paying less for those impressions, as I learned when my impressions dropped slowly to nothing over the course of January.

If this setup works though, it will be ideal. I will be paying exaclty how much I want to be spending on avertising, and it will be impossible for any fraudulent clicks to occur with this sort of campaign. Some advertisers may choose not to run my ads because they're not resulting in profits for them, but those same campaigns wouldn't result in profits for me either, and I'd actually be taking a loss running in those places.

So I'll let this new setup run a couple weeks and see how it works. Hopefully sales improve as a result. Or at least, my profits.

At least you've never had a month of no sales, I had one of those with my framework and it was sad :-( Most I've ever sold in a day is 3-4, most sales are in January (new year, new goal to make games). Sounds like we've made similar profits (mine's about $8k too and I started selling in May 2005 (wow nearly 2 years!))

Grey:
I've had plenty of months of no sales of my various systems. Those only sell maybe one or two copies a month at most. The market for Blitz users is much smaller than the marker for 3D artists. Sounds like your framework is selling better than most of my stuff. (Though my sprite system for BlitzMax sells a couple copies a month on average.)

I could give you some tips Swift that could help you boost your sales.

If I was you I would spend some of that hard earned money making another product that has to do with textures. I would make a texture CD or texture pack that is downloadable. You clearly have a talent to produce that.

Then I would put up a campain and announce on every forum and community where there are potential customers that I am lowering the price of STG for 10 days, in this frame the price would be 10$ cheaper. (Since you say you are willing to give 15$ per sale, 10$ cheaper is nothing)
After the 10 days run out the price would go back to normal. You could also set this up so that each day price would go up with 1$ so that customers rush in a buy at day 1 so that they get it cheaper.

Even if your software is cheap, people are even cheaper and if you give them rabate they will think I better buy this now even if I really dont need it I may need it later, why should I pay 35$ later when I can get it for 20-25$ if I buy now.


We Tridinaut did this once and in 1 week we made 1000$ worth of sales. We did not even advertise with google, all that we did was post on perhaps 5-6 big forums where I know game developers hang out. That was enough. They saw the opertunity and sales were going creazy.

Now that is one tip that you can use that can make you money without shelling out any money to google ;)

Also as I mentioned you need more then one product if you are going to survive. You then use your email list of customers that have already been with you and announce these new products or announce that you are running a campain where price is lower for 5-10 days.

Now that way you can start making money without having to invest lots of money into google. Build up a store and do not belive that you can make any living with one product, there are not enough customers for that.

Dont waste your time and money into investing into google so much, invest in some new products and build a customer base and then market all new products directly to people who you know are not afraid to open their wallet ;)

If you are interested into making more money I will give you 50% on any sales you make of existing tridinaut products. That way you can offer your customers more products from the start while you work on making more interesting offerings. You need to offer more then your little magic program.

You need to offer a bunch of good valuble things. If you dont have time to produce it your self, then partner up with other authors. Find someone who is selling good texture packs and ask them if they are looking for affiliates. If you are ever going to make lots of money you need more product either your or you partner up with other people. That's how money is made.

Maximo:
Acutally I already have a texture CD made. It's not a great one since the photos were taken with an older camera, and the vast majority aren't seamless. But it does have 1200 images, and I've been thinking about offering it for free with the texture generator to provide an extra incentive to purchase. I just haven't gotten around to it.

And thanks for the offer, but I'm just gonan stick with selling my texture generator for now. I'm not in the poor house or anything and I really need to concentrate more on getting my game done and less on improvidng sales of a product which I can't makke a living off of even if I double sales. :-)

Do you use Google analytics?

No. I'm not sure how that would help me tell where my sales are coming from, since the visitor logs on my website just give the google ad server as the refererer. ALso google already tells me on which sites users clicked on my ads.

Except for the pay per download campaign, when I run that. But knowing which sites users click on for that isn't really useful.

What I need to know is which clicks ultimately result in sales. I've changed my pay per download campaign to a pay per sale campaign for the time being because at least that way I know I'm only paying for ads that result in clicks. But I don't get near as many impressions with that setup. But maybe that's simply because my program can't support more ads than that.

I have been meaning to look into analytics though anyway to see what it can do. I just have my doubts it will be useful to me.

Analytics gives you a lot of nice statistics. Not sure it can really tell you who clicked what page, unless it's pages within your own site.

You need to get the demo of seamless texture generator included with all manner of game dev tools.

I am sure you have though of all pricing models, but perhaps a free version that is limited to generating 256x256 size images, rather than a time limited demo?

Maybe. Sold 17 in Jan and 12 in Feb. That was a pretty good "Haul". One more sale and I'll hit 150 paid up users. I'll tell you what REALLY helps my sales: releasing games made with the framework. Every time I release a game I get an influx of sales and Fairway Solitaire helped the sales a lot. If only you could hurry up and release a new game using your sprite engine...

I am by no means an expert, nor in a qualified to give an expert opinion and havent released anything exept my imagination :P

BUT, that sounds nice, and is good for a promotion to get your products and word out. Sswift would need to sell (i guess) 10 copies at promotion-price and therfore actually loose out on the deal and the nett effect is no $$-gain. To relative, car dealers never/seldom sell their old models at a reduced price just coz a new model lurks around the corner.

Maybe sswift can add some small tidbits like extra foto's and whatnot for a small microtransactions of $1-download price?

Sswift thanx for your insightfull and very informative post :)).

The demo isn't time limited. The demo puts watermarks on the images it generates and prevents you from saving.

A demo which could generate only 256x256 images wouldn't sell, because people would assume that's as high as it can go, and/or would want to know what the textures look like at high res. Nobody wants to make 256x256 textures anymore.

Have you approached Garage Games about publishing? Looks like The Game Creators have their own texture maker. What about the Unity guys? Surely piggybacking with these type of operations is you best bet.