Home > Paid Traffic Sources > eCommerce

---

White Hat Beauty on Amazon / Paid - Spent $690 on Facebook -> got 1 sale (38)

---

Hi STM

Hope everyone is enjoying their weekend

After a fair bit of trial and error, I decided to create a follow along thread to get some help. I have started building the funnel, but it does not seem to stick in terms of paid traffic so far.

I picked female beauty because I have run skin free trials for many years with great success. When I first started doing free trials and was making a nice profit, I started building out a long term blog focusing on SEO traffic. This site now does 2k+ traffic per day from SEO. I then took this site and created a brand around it. The store has been live for almost 1 year on Shopify and Amazon FBA (fulfilling Shopify through Amazon).

Currently 5 products.

Sales on Amazon are steady 10+ per day. Reviews and feedback from customers are great.

Shopify store is using Retina with apps like Abandon Cart, OCU, Shopify Reviews, Retarget App ++.

Soo, we are trying to crack the paid traffic for this store. Decided to start with Facebook.

Here are the stats so far (camp is paused as of now):

Imps: 18747
Clicks: 484
CTR: 2.58%
CPC: $1.43
Cost: $690
Add to Cart: 10
Purchase: 1
Revenue: $25.66

Bid: Auto CPM
Objective is Offsite Conversions using Add to Cart as target.
For interests I have tested blog visitors last 180 days (from SEO), shop visitors, lookalikes and people that like competitor brands.

Have tested using a pre-sell LP (5 tips why xxx helps xxx) which links to product page mostly.

LP CTR is 32%. So people are clicking through on the LP. 10 are adding to cart. But only 1 ended up purchasing.

Have tested multiple ads.

Using Adespresso to setup the campaigns and Voluum to track.

This campaign is now paused because it's done really bad!

Right now I am doing a image test, as iamattila suggests in Part #1 in his post.

8 images (1 ad each ad group, total 8 ad groups)
$35 / day budget
Auto CPM, Objective Offsite Conversions -> Purchase
It's run for 48h now so I will pick a winner later today, will update on the stats as soon as their in.
Target Audience: Blog Visitors last 180 days

But yeah, overall kind of flying a bit blind here.

Would be awesome to get some help from the Ecommerce / Facebook experts on STM I am open to test anything!!

It may be that my product page sucks. Using traditional page, haven't tested long form yet.

Thanks for your time and have a great day!

// Stian

---

Good on you for trying to tackle FB, it's not easy initially but can be super profitable once you find the sweet spots.

I would look into split-testing interests, they're key, but may require some extra market research on your end. You already have some data so you'd know where to start. Targeting blog visitors might sound like a good idea but if people don't buy anything I'd turn to retargeting your shop's visitors instead with some clever angles. Aside from those, try and research extra relevant interests with a higher median number of ads clicked (in Audience Insights) and split them in separate ad groups/campaigns. Don't keep your targeting/audiences too broad nor too narrow while doing this.

It's kind of hard to tell without seeing what your landers or products look like, but it might be worth testing direct linking to your product pages (if they're well optimized). Another bit that should most definitely be addressed is doing some conversion optimization on your site - often times this is the one thing that can turn your campaigns on the profitable end. Look into completely destroying any objections that your customers might have during the entire checkout process - I.e. tell them that shipping is free, the checkout is secure (with the obvious secure badges), let them know they're being well taken care of all the way until they hit 'Checkout'. Add the option of being able to checkout directly and skipping the whole add to cart process with stuff like PayPal Express checkout etc.. Provide upsells/downsells/cross-sells if you have extra products, give your customers discounts.

And finally, test as many angles as you can, as mentioned in Attila's post.

All in all, look into testing every possible "setting" on Facebook - be it ad types, ad objectives, conversion events, various interests, bidding modes (manual vs oCPM etc.) age groups, angles(and your angles' copy), placements, images, video, product ads etc.

Split your ads into age groups to find your most profitable ones, coupled with relevant interests and good ad images/videos, they can get you some high relevance scores.

If you have an email list of your customers, make a lookalike based on it.

I usually start some pretty broad tests and gradually optimize from there, it's more expensive but worth it in the end. Test tons of images, split by age group, 2-3 angles per campaign, and then have a bunch of campaigns each with a couple of their own interests and the same campaign structure. This allows you to see your best images/angles/age groups/placements and gives an idea on where to drill-down on interests. This can result in possibly hundreds of ads being run at the same time and a pretty high cost, but the insights you gain can be superb. You can also do smaller budget tests, and do it step by step too, everyone has their own testing methods.

Looking forward to more updates!

---

If you have a fan page, you can start making it popular first. If you don't have one,you should start one.

You can make a post about your best-selling product with some angles and boost the post with post engagement instead of conversion. Then "website click" and finally use "conversion".

---

@stian I would seriously look at cart abandonment. Why do 9 people leave without buying?

If you want to share the product page / site, I can have a look at it.

A pro-tip here: add chat to your site. Even if you don;t intend to have a live operator. Just had a customer today leave a message saying he's a registered business and wanted to buy excl. VAT.

Left his e-mail in the off-line message, so we can follow that up tomorrow and rescue the sale.

And there are a lot more people leaving messages, so it's well worth setting up.

We use clickdesk(.com), but there are plenty of options.

---

Thank you, good tip. We have a fan page already which has a decent number of likes. Would

Would you say it is better to boost a page post than to just create a "normal" ad?

Thanks so much! And thanks for the PM tips Yes, we have implemented a chat.

Stian

---

Thanks for your awesome reply shishev!!

With so many "variables" when it comes to FB, I think it needs to be structured in a way (unless you wanna do the bomb approach in your last paragraph .. )

Here's what I've come up with, let me know yours thoughts (especially if we should tests angles or interests first):

In order first to last:

1. Finding your winning image (iAmAtilla's post)
2. Got the highest CTR (All) image? Good, now the fun begins - Test 10 angles - $40 per angle (iAmAtilla's post)
3. Interests / Custom Audiences - based on RT lists and higher median number of ads clicked (in Audience Insights)
4. Ad Types
5. Ad Objectives
6. Conversion Events
7. Bidding Modes (manual vs oCPM etc.)
8. Age Groups
9. Placements

I am sure many of these can be tested at the same time, but this is just a rough draft.

So far I have completed number 1.

Stian

---

If I understand correctly, you use your website traffic and feed the FB Pixel.
But then you used ATC as a conversion target ?
You ask the FB pixel to optimize for conversions with people who added to Cart, but the pixel does not have this data yet.

At, least this is my understanding, external data are awesome to create lookalike audiences, but if you want to use the different optimization point you need to feed the beast...

---

great thread btw :-)

---

Yup, I'd agree with Pekadis - the big red flag in the numbers for me was the 90% cart abandonment rate.

I'd recommend looking into that as top priority, figuring out what's going on, and probably implementing some cart abandonment reclamation software /routines.

Looking forward to your next results on this - interesting thread!

---

Are you running PPE or WC? Additionally, as mentioned above, you need to figure out why your cart abandonment is so high. Optimizing your store is half of the process. Use a longform funnel.

---

Hi, the same FB pixel is implemented on both the blog and the shop. So it has been recording all sales data Shopify feeds it for the last year (there has been quite a few organic sales). I am trying to feed the beast as best as I can, but beast aint giving me anything back

---

Thanks mate for your reply! I have bought the Abandonment Protector App - which sends out emails after a customer abandons cart. But this only works if they have started checkout.

When a customer clicks "Add to Cart" the product is added to cart, and they are redirected to the /cart page - this is where the drop off happens, they don't even start checking out. So I will take your advice and make this the #1 priority

Thank you pain2k. Running WC with Add to Cart as target. I have Hotjar and Optimizely setup =)

---

I had a chance to take a look at @stian's product page and wrote out some quick feedback, I'll paste it below so others could learn too. Without revealing the website, the product page is what you would imagine a typical product page, nothing too special about it:

- Test having a Buy Now/Checkout button instead of the add to cart button. Or maybe even both.

- Add a discount on the price. Doesn't mean you'd have to change the price tag but maybe go for something like $XX.99 -> $XX.99, so at least you're showing a discount.

- Add security badges - https://baymard.com/blog/site-seal-trust Check here to see which ones are the most trusted. I.e. Norton, McAfee, TRUSTe etc.

- Add some urgency if you can. Say, coupled with the discount add a timer + limited quantities: "XX% OFF, Limited Quantities, Will Run Out In [INSERT TIMER HERE]". Possibly something a bit better worded and mroe creative.

- Perhaps having some more images and (if you have it) a video of some sort. Might be worth paying someone to do a quick review of your product too, which is something you can later test on FB as well.

- And finally, if you decide to only go with "Add To Cart", test out different cart pages. Yours currently looks a bit empty, and not very trustworthy. Could be the reason why you have such a high cart abandonment rate. Get some trust badges in there. Add an exit intent pop-up on the cart page with a nice discount - "Get 10-15% OFF your checkout, enter your email" type of thing.

Once the above has been done, study your analytics and see if the bounce rates have decreased and conversions increased etc. I'm pretty confident that they will.

As for cart abandonment, I quite like Ghostmonitor (now named Recart https://apps.Shopify.com/ghostmonitor), seeing some very good results from it lately. You can also squeeze out quite a few days for a free trial if you'd like to just test it out against your current app/plugin.

I'd also highly recommend that you look into conversion ads if you're looking to test straight sales (product page).

---

Ho, so if you have already a good amount of conversions, why do you target ATC ? FB is actually giving you what you asked for.
I would totally just duplicate this campaign and change Conversion point to 'Purchase' ... that the ultimate goal, right ?

everything show up here ? : https://business.facebook.com/ads/ma...acebook_pixel/

Of course, optimizing your shop is great but you said this :

---

Thanks for your help mate!

I have followed your advice and implemented the following:

- Test having a Buy Now/Checkout button instead of the add to cart button. Or maybe even both.
-- Running Buy Now button only now which takes user straight to checkout

- Add a discount on the price. Doesn't mean you'd have to change the price tag but maybe go for something like $XX.99 -> $XX.99, so at least you're showing a discount.
-- Added

- Add security badges - https://baymard.com/blog/site-seal-trust Check here to see which ones are the most trusted. I.e. Norton, McAfee, TRUSTe etc.
-- Added below "Buy Now" button, HQ images

- Add some urgency if you can. Say, coupled with the discount add a timer + limited quantities: "XX% OFF, Limited Quantities, Will Run Out In [INSERT TIMER HERE]". Possibly something a bit better worded and mroe creative.
-- Have not added this yet, but have looked into getting Hurrify

- Perhaps having some more images and (if you have it) a video of some sort. Might be worth paying someone to do a quick review of your product too, which is something you can later test on FB as well.
-- Added to the todo list

- And finally, if you decide to only go with "Add To Cart", test out different cart pages. Yours currently looks a bit empty, and not very trustworthy. Could be the reason why you have such a high cart abandonment rate. Get some trust badges in there. Add an exit intent pop-up on the cart page with a nice discount - "Get 10-15% OFF your checkout, enter your email" type of thing.
-- Running straight to Checkout now, but will improve cart as well for future use

I'd also highly recommend that you look into conversion ads if you're looking to test straight sales (product page).
Yep, been running conversion ads straight to Product Pages (results in next post).

Thanks again for your awesome advice, if there is anything else you can think of, please keep them coming

Stian

---

Hello

Update 12. March

Results from Step 1 - Image Test

Overall

Sales: 0
Spent: $66.09
Objective: Website Conversions with Purchase Goal
Bid: Auto CPM
Targeting: 1 interest
ROI: -100%
Impressions: 3606
Link Clicks: 64
CTR (Link): 1.77%

I used CTR (ALL) from FB stats to measure this, not CTR Link Clicks (hope this is correct).

Impressions - Clicks (All) - CTR (All)

1: 475 - 9 - 1.89%
2: 453 - 7 - 1.55%
3: 519 - 11 - 2.12%
4: 458 - 11 - 2.40%
5: 517 - 9 - 1.74%
6: 349 - 8 - 2.29%
7: 419 - 13 - 3.10%
8: 416 - 12 - 2.88%

Decided to go with 7.

Results from Step 2 - Angles

Tested 10 different angles I bought from Angle Saurus. I have strong tech skills, but copy writing has never been my strongest side, just takes ages and gives me a headache each time! So decided to outsource it. Here's the stats:

Overall

Sales: 2 (Woop!)
Spent: $359.14 ($36 per angle)
Objective: Website Conversions with Purchase Goal
Bid: Auto CPM
Targeting: 1 interest
ROI: -85%
Impressions: 11779
Link Clicks: 487
CTR (Link): 4.134%

Click (Link) | CTR (Link) | Sales | ROI
1: 59 | 5.099% | 0 | -100%
2: 39 | 3.757% | 0 | -100%
3: 63 | 4.415% | 0 | -100%
4: 24 | 3.154% | 0 | -100%
5: 43 | 3.993% | 0 | -100%
6: 15 | 1.674% | 0 | -100%
7: 52 | 4.887% | 0 | -100%
8: 56 | 3.862% | 1 | -55%
9: 68 | 4.558% | 1 | -55%
10: 68 | 4.799% | 0 | -100%

Finally got some sales and we have 2 angles that shows promise!

Next step now would be to test "3. Interests / Custom Audiences - based on RT lists and higher median number of ads clicked (in Audience Insights)"

Does the testing look ok? Settings etc? Any input from you guys is much appreciated

It still overall at a massive loss, but guess I will get there in the end !

Have a great Sunday!

Stian

---

One tip for you guys using Shopify with Voluum; here is the code I use under Settings > Checkout > Additional Scripts to postback conversions with total revenue to Voluum.

Only triggers once per customers and sends full revenue back (inc tax / shipping).

landing_site_ref is sent from Offer in Voluum like this: http://myshopifystore.com?ref={clickid}

Best would be if we could do some calculations with "total_price", so that we get the actual "profit" sent back to voluum. Any ideas?

{% if first_time_accessed %}
<!-- Conversion scripts you want to run only once -->
<script type="text/javascript" src="http://trackerurl.com/conversion.js?cid={{ landing_site_ref }}&payout={{ total_price | money_without_currency }}&txid={{ order_number }}"/>
{% endif %}

Stian

---

That should be perfectly doable in Liquid, although you'd have to hardcode your costs into the source code or the Shopify database - and be aware that means that users could find out your costs and profit margin if they go snooping!

---

Based on 2 conversions over 10 groups, I don't think you can say that 2 angles are showing promise.
I am not an expert in statistics, but I don't think you have enough conversions to call a winner. It's just coincidence.

Plainly put - I think you just found the right 2 customers, that would have bought anyway, regardless of the angle. And you don't have the numbers to prove me wrong.
Not trying to be a d*** here - just would like to prevent you from jumping to conclusions.

Still surprisingly low conversion rate.

How are the conversion rates compared to the SEO traffic? With 2K per day, you'd be able to test more than with your ads. And there are no traffic acquisition cost involved.

My gut feeling is that the offer just isn't good enough.
Could be the price perception, images, copywriting, benefits etc.

But before I jump to conclusions, it would be good to compare conversions on organic traffic to ad traffic.
If both are low, it's likely not the angle (meta description would just give you one angle).

And from my copywriting experience, if it doesn;t convert, the easiest fix is the offer.

Hope this can help in turning things around.

---

Thank you.

---

Since you have just 10 people adding products to the cart and just 1 sale, why don't you call/email those 9 dropoffs yourself and ask them why they dropped off? You'd be surprised by the honest feedback people give you.

---

Thanks for your reply!

Yes, it may just be that my product page sucks lol. Based on your feedback, I guess I should run the campaign some more while split testing the copy of the product page. Even tho we are closing in at around $40 per angle.

CR% for SEO traffic is 0.94% so far this year, for Angle FB Campaign it would work out to 0.41%. So both are low.

So far using traditional product page, kind of same as Amazon does. Image on the left, bulletts/reviews/buy now on the right, and more copy below. But guess I should look at a long form one.

Appreciate your input

Stian

---

Hey Stian,

I have been going over this thread and checked out the "Abandonment Protector" you recommended. While going over the reviews there was a really bad one (actually the only bad one) saying they guy there had a drop in CR when he used their "contact Us" pop up.

This is a part of his really looonnng review:

"I was running facebook and other promotions and getting very small abandon carts, yesterday I had 105 visitors and only 2 person added to cart. I kept thinking that it was the campaigns and kept killing them. Then I added a recording app so I could see where customers were dropping off. Only to see that when visitors land on the page,THE CONTACT US POP UP CAME ON AND COVERED MY SALES PAGES AND THE VISITORS HAD A DIFFICULT TIME GETTING RID OF IT"

Not saying that app is good or bad (the reviews actually suggest its a good one) but thought it might be worth while checking you don't have a similar problem...

---

Hey STM

Small update. I decided to continue the angle test based on your feedback. But it's taking ages to get statistical significance.

Here's the stats so far:



Let me know what you think

Stian

---

Quick update, paused some of the angles yesterday. Now left with 5 angles that have 1 or more sales.

CR & CV is low.

Frequency & CPC cost is climbing. Freq = 1.32, CPC closing in on $1.

Audience is 500k+ may not be big enough.

---

Thanks mate. I have quite a few that abandon cart. I just activated ReCart as shishev recommends. Will let you know results vs Abandonment Protector

---

Nice - progress being made!

What's the LTV (Lifetime Value) of a customer for you at the moment? If it's genuinely just the revenue from a single sale I'd say that's probably one of the easiest places to improve your ROI. Can you upsell, get them on a list, or similar?

It might also be worth trying a lead capture -> email sequence -> sale funnel rather than a straight ad->purchase funnel. On larger-cost items, capturing leads then running an email sequence with genuinely useful info that offers multiple conversion points will often have a higher ROI than a straight-up "buy now" approach.

Have you had a chance to set up a split-test of the lander for your product yet? That's the other area where big gains can probably be made comparatively quickly. If you fancied sharing the lander with STM, we can probably give you some suggestions.

---

Thanks caurmen!

I used this article to calculate LTV a few days ago, it's $21.40 so far which is mostly revenue from a single sale. Store has a total of 6 products. I have upsell setup during first purchase, but not after a customer has purchased, guess I should work on an email sequence for that. As mentioned before, I am not a strong copy writer .. Does anyone have recommendations for where I can find a freelancer to write those types of selling emails / email sequence?

lead capture -> email sequence -> sale funnel is on the list also.

I am split testing a few tweaks using Optimizely. But yes, I think my product page has great potential. I don't wanna share the link in public here, but I can PM it to you - if you wanna take a look?

Thanks again for great feedback.

Stian

---

Hey, I wanna share some thoughts on this. Lets look at the math:

"CR% for SEO traffic is 0.94% so far this year, for Angle FB Campaign it would work out to 0.41%. So both are low."

Lets assume with a better funnel, FB traffic converts like organic SEO traffic. about 1%, with average value about $25. Lets ignore the future monetization of the customer for now... in order to break even on the front end with a 1% CR you need to clicks for $.25 CPC. If your clicks are closing in on $1 CPC right now its gonna be tough, even with amazing on page optimization and sales copy.

What I would do is get a facebook ad spy and look at ads in your vertical. anything doing well on there is worth testing. But be careful, a lot of them will be aggressive so understand there might be a little risk for your account (tone the copy down if needed). For landing pages you'll see a lot of cloaked uncompliant stuff, take the general ideas of those and create your own compliant versions, use them in your campaign flows. By doing this you will be modeling successful affiliate campaigns in your vertical to promote your own stuff.

If you get your cost down and conversion rate up (and copying whats working is a way) you'll be able to buy more traffic and use all that other optimization to improve the ROI from there. But if you keep paying $1 per click figuring out the value of your optins from a split test or testing copy to stop shopping cart abandonment, its gonna cost a lot to get significant data.

---

You may struggle to find a really good copy writer unless you have quite a high budget - it's an in demand skill set.

However, it's also very learnable. That may be an alternate way to go - grab and read Cashvertising, Words That Sell, Tested Advertising Methods , read through Copy Hackers, and you'll probably find you can write something that converts more strongly. As mentioned above, spy tools are also very useful.

---

Update 29/3

Ok, after $2131 spend on this campaign, I had to pause it.

CPCs were rising and conversions stopped coming.

CPCs seem to have gone above $1.1 after March 20th, even as high as $1.4, $1.8 on some days. I probably paused some ad sets around that time, but cannot remember doing anything else.

Relevance Score went below 8 on the 20th tho. Was 8.1 from the start, but then went down to 7.8 -> 7.7 -> 7.6 and kept that level.

Frequence went from 1.06 on March 8th -> 1.5 March 28th.

I paused a few ad groups on Friday, and evened out the total budget across the active ad groups. After I did that I havent got any conversions (may have been coinsidence).

Here's the final stats for this campaign:



I am not sure what the next steps are, have to think a bit about it. I guess if anything, I could take Angle ID 1 and 9 and work with those, but still very low RPC/ROI.

Getting the product page redesigned, hopefully that will help increase conversions.

Right now a little depressed this didnt work out at all. But ohh well, guess I will get there in the end

Thanks for your suppor everyone! Any tips on next plan of action is much welcome!

Stian

---

Quick question to the FB experts out there; what bid mode do you recommend?

For the campaign in this follow along I have been using Pay for Impressions with Automatic Bid.

I see now that I have been paying between $30 to $46 CPM, isn't that insanely high?

At $30 CPM, you need a CTR of 6% to achieve $0.5 CPC.

At $45 CPM, you need a CTR of 9% to achieve $0.5 CPC.

Is it really this expensive or am I doing something very wrong?

Stian

---

Another question for the FB experts

Do you recommend creating a new advert for advert purpose only? Or create a new page post / use an existing page post from the brand fan page and boost that one?

---

If I were in you, I would have split that ID9 and get it profitable. Sometimes it takes so little to get from -60% to (at least) break even.

Depends on your objective and your brand.

By the way, I'm not quite expert in manual bidding, but I don't think your calculations are right.

---

Thanks for your reply mate.

In regards to objective and brand; we are trying to build a serious, stable and long term brand. With this extra info, what would you recommend for my original question?

I used this calculator to get those numbers. Just tried again and I am getting same result .. may be wrong tho.

---

Small update!

So after following sapven's thread, I started testing Franklin Hatchett's methods for FB Campaigns. Here is my breakdown of his method:

Source: https://www.youtube.com/watch?v=zxpiAaVfGXc

1. Run PPE advert until 2.6x offer payout
2. Take the best advert from PPE and create WC campaign with same ad, auto bid, using VC as target
3. Let advert run to 2x offer payout
4. 0-1 sale, keep it running
5. Let run til 3x offer payout;
a. 0-1 sale but same price: kill it
b. Or if still 1 sale, but cheaper actions, let it run for 1x offer payout more (total 4x offer payout)
c. Or 2 sales at 3x offer payout: keep the advert running
6. If total spend $150-200 and not profit > $1: kill it

My store has a total of 6 products.

I have tested the method on 4 of the products, 2 which are live at the moment.

Product 1 got 125 View Content and 0 sales, spent 4x offer payout.

Product 2 got 155 View Content and 2 sales, spent $160 which is more than point 6, so killed it.

Product 3 & 4 is still testing.

So Franklin's idea is if your product is good enough, it should get sales when targeting VC.

You need to supply the FB Pixel with 15-25 conversions per week per ad set for the WC to work properly. And that is much easier with VC than using Purchase.

I will continue testing Any input is much appreciated !

- Stian

---

That seems like good advice to me.

---

Great thread! props for getting strated with paid after cashing in with SEO, it's a tough move.
Im actually intereted to know about -

Is your organic traffic comes mostly to product pages?
Arent you afraid of making a lot of changes to these pages in terms of affecting the on-page seo 'magic' ?

Also, were you affected at all since the latest updates kicked in (fred etc..) lots of turmoil!

---

Home > Paid Traffic Sources > eCommerce