we are doing “one-sales desktop” software. I think there are very few companies who actually make money in this segment through Facebook advertising. You see little desktop software advertising on FB… Retail clothing etc. seems to be great in terms of ROI. The problems in my experience:
- Conversion rate is generally low in software
- Many people focus on website clicks only - but cheap ones show little conversion
- It is hard/impossible to accurately understand what lead to a sale
I think Facebook itself is the most powerful advertising channel that can generate LASTING revenue. However, no sales channel works on its own. People need to hear from you through different channels:
- Get bombed with Google image ads
- Other relevant channels
You need all of them in our experience. Also: “Crossing the chasm” is true. Also: software needs about version 3 to cross the chasm ( Joel was simply right as well ). It takes time. We do spend a larger amount on FB (couple of thousand dollars per month) - for one product. The biggest challenge for software sales in our opinion is: you won’t find out how people found out about you and decided to buy from you / there are more than one channel responsible.
Tracking sales accurately is essentially impossible. Google Analytics records all transactions associated with a user flow for 4h if he is constantly busy on your page. So if a user sees your ad on Facebook, visits your homepage, downloads your software and tries it out, has lunch, returns to his desk and decides to purchase it will show as “don’t know where he came from”. With free trials covering several days/weeks that is even more complicated. Associating sales to a particular advertising campaign is essentially impossible. Also considering people may click multiple ads (e.g. Facebook, Google)
What we found to work for us is to focus on “downloads” as a metric for advertising spend. People decide within minutes whether they download and try our software - so it is rather easy to determine the “cost/download” for each campaign. Every Facebook/Google campaign should have it’s own URL (URL Builder). This way you can improve performance drastically. I think if you spend a couple of thousand dollars on FB, you will see a decent ROI. The question is: how can you spend 1, 5, 10k every month and have a 300, 500, 1000% ROI for longer periods… That’s a tough one. I would say that in the beginning (when relatively few people know your software and you may not exactly know your market) up to 50% advertising spend (% of revenue) is acceptable, but it should decrease over time to about 20-30% (20% would be really good in software IMHO). Also: it really matters where you are in the life cycle. Crossing the Chasm is simply amazing as a book.
How successful campaigns worked for us:
- Upload customer list on FB, create custom audiences, 2% size (3+% is too big) - but you need many customers
- Have high quality image material
- Have good text
- Change text / images regularly - e.g. once every 3-4 months
- Try different campaigns, but always start a new one, never edit existing ones
- Age, gender, country and interests are (at least for us) the most important success factors (if you don’t have enough customers yet)
- Delete negative comments immediately (ideally blacklist all comments)
- Be rigorous in deleting non-performing campaigns, but spend at least 250+ USD per campaign
- REMARKETING is big on Facebook - you should do it
- If you optimize for conversions, Facebook can achieve OUTSTANDING results consistently - e.g. .2$ ad spend on the 1$ of revenue… However, those are very conservative, they will use little budget and probably not boost your sales a lot…
- Optimizing “page visits” and verifying the campaign through Google Analytics with “costs per download” worked best for us…
With Google its more difficult. We had little to no success with the ads leading sales… Image ads, Remarketing and building up reputation did work. Having Google optimize Ads for downloads etc. never really worked for us… Bing is just a waste of money… .05$ per click - but not a single DOWNLOAD…
It is an interesting topic, unfortunately there is too little good material out there. Most blogs talk non-sense a lot. A/B split tests also make very limited sense in many cases - the statistics are just wrong. There is too much variation.
On another note: we only look at sales on a weekly basis - also because our customers come from all over the world. The only thing that boosted our “download per page visit” conversion was putting the download link on the first page. Everything else (even if A/B tools showed that other changes we did were awesome) just “faded away over time”. There is too little reliable information - unless you are Google or Microsoft or Apple. Don’t trust mathematical models / statistics in small software companies. Data/noise ratio is bad.
So yeah: FB can work very nicely and sustainably for us, 20-30% ad spend would be great in the segment, scaling is hard, tracking sales is even harder - what are your suggestions?