Table of ContentsThe Opportunity
Pinterest still has a poor reputation with some advertisers, and it’s deserved… to a point. I was a Pinterest early adopter. 5 years ago I was one of the first to ever run Pinterest Ads, back when:
- Stats didn’t update until 24 hours later
- You had to use a software like 4C to even upload ads
- You had to have dozens of ad groups and ads as well as bid for clicks on everything
On top of that, the volume used to be limited… Now, all of that has changed.
Before the IPO (and after), Pinterest has made a ton of improvements to its ad tech. New additions in the last year or so:
- video ads
- collection ads
- conversion bidding (oCPM)
- automatic bidding (in beta)
- dynamic retargeting
- dynamic prospecting
- offline sales attribution
And on top of that, there is more available inventory (impressions) than ever. Pinterest has experienced massive user growth post-crisis, going from 250M to 450M monthly active users. And the icing on the cake is, over time the demographic on the platform is shifting older, and more male, which opens up more opportunities for advertisers.
Pew Research puts out surveys on social media use every year, and as you can see below, more Gen X and Boomers are using the platform than ever before. And as we all know, those are the demographics with money to buy your products.
Pinterest Ads for E-Commerce
The big key to success for e-commerce brands on Pinterest is making sure you run dynamic campaigns (shopping objective) and conversion objective concurrently. Pinterest reps shared with me that conversion rates for users who see both a single image or video ad, and then a dynamic product ad, those CVRs are 6x higher.
So the first thing you need to do is upload your feed into Pinterest, and make sure your feed is up to spec.
The great added bonus about uploading your feed into Pinterest is that once you do, all your products will be surfaced organically in relevant searches. This means you get a ton of FREE traffic. (Depending on how many products you sell, I’ve seen companies get 10-50k a month in free traffic just by doing this).
OK, so your feed. Make sure every single product has a Google product category. Pinterest, unlike Facebook, won’t accept any SKUs without a category. And they also use this info to inform the targeting of shopping campaigns. So make sure your categories are 4 or 5 levels deep as well.
Want to avoid manually editing every row of your feed in excel? A quick “hack” I’ve used in the past is uploading your feed into google merchant center, then downloading it again. Google will automatically categorize every product uploaded into merchant center, and you can use their categorization for your Pinterest feed!Account Structure
Ok so you have your feed ready to go, now it’s time to get some campaigns up. Assuming you sell more than one product, you want to set up a full-funnel structure using both conversion objective and shopping objective.
This means the bare minimum you should have running is:
- one shopping prospecting campaign (Pinterest DPA),
- one static prospecting campaign (conversion objective),
- One shopping pre-purchase retargeting campaign
- One conversion objective (or traffic) pre-purchase retargeting campaign
- One shopping retention campaign
- One conversion objective retention campaign
This would be for a low-medium traffic brand. As you achieve more scale you can add experiment with adding another pre-purchase retargeting campaign that is focused on cross-sell.
For shopping prospecting, an easy way to start is having one ad group per category or subcategory in your assortment. Most of the purchase intent on Pinterest is category driven so we don’t want to get much more narrow than that to start.
Once shopping is running, experiment with different advertised product groups like:
- Bestsellers (overall)
- Bestsellers from Pinterest (can be gleaned from google analytics or Facebook analytics)
- Trending products (sales volume has increased over X% MoM)
- Products high LTV buyers tend to buy first
For your static prospecting campaign (conversion objective), pick your 2-3 highest selling categories and start there. We’ll touch on this in the next section, but the best targeting you can do here is keywords so I’d start with 5-10 broad match keywords per ad group. You can get these from your Pinterest rep, or from Pinterest Trend tool, or from google keywords planner!
For your pre-purchase retargeting campaigns, the thing to keep in mind is balancing audience size with bid type and messaging. For e-commerce, we know intent to buy tends to drop off 3 days after a key action (PDP page view), and then again after 7 days, then again precipitously after 14 days. For higher-priced items, the cycle might be 7 days, 14 days and 30 days.
Considering Pinterest’s delayed purchase cycle (pins users take longer to buy), let’s start with the following segments:
- 0-7 PDP page viewers, excluding add to cart
- 8-30 PDP page viewers, excluding atc
- 0-7 atc, excluding purchase
- 8-30 atc, excluding purchase
Depending on how much traffic you get, you’ll want to experiment with bidding for conversions on these segments vs bidding for impressions or clicks. If any of these audiences is less than 100k a click bid is a good bet. If it’s less than 20k, impressions will probably get you where you need to go.
Unlike Facebook, Pinterest doesn’t aggressive segment their users into buckets. So you can still get purchases by bidding for something other than purchases.
For your retention campaigns, a good start is two ad groups for the shopping version. One targeting purchasers 0-30, advertising all products. And one targeting purchasers 31-365, also advertising all products.
The idea here is that Pinterest’s taste graph (what feeds shopping targeting), will automatically serve the appropriate products in your assortment to someone who has purchased from you before, based on what their Pinterest activity has been since that purchase.
For the static retention campaign you can experiment with cross-sell, but you may have to expand your lookback windows a bit. E.G. Purchase 0-4 months ago of best selling product 1, show them best selling product 2, and vice versa.
The static campaign will allow you to customize creative and copy more, so don’t be afraid to offer coupons to buyers in that campaign to incentivize the second purchase.
The advanced version of these retention strategies would only be advertising to users you are confident are past the time period where they would normally make a second purchase. You can glean this information from FB analytics or google analytics cohorts. Look at where the cumulative revenue per user drops off and start there.Targeting Hierarchy (In Order of Profitability)
So when you want to scale your Pinterest account horizontally, here is the order in which you should think about adding targeting (in order of ROI):
- More shopping ad groups
Expanding your shopping targeting is really self-explanatory. Add another ad group and another product group that corresponds to a subcategory in your assortment, and you’re done. In a perfect world you have one ad group and corresponding product group for every subcat in your feed. And you control for cost and varying returns on each product by setting manual oCPM bids.
Keywords are going to be your best static targeting on Pinterest. And you can imagine why. These are people who are searching for products, services or topics related to what you’re selling. Keywords on Pinterest are “broad match” by default, but you can set an exact match via spreadsheet upload (don’t do that).
Beyond looking at google keyword planner and asking your Pinterest rep for trending searches, you can also leverage Pinterest trends tool and Pivot off their related searches. Say you’re selling makeup, and you input makeup routines:
They will give you related terms at the bottom which you can click on, which will then give you another set of related keywords to target.
Like Google, there is an element of relevancy that determines what ads serve on what search terms. So make sure to group your keywords into relevant content groups, and have ad creative and copy to match those searches.
Pinterest also has a highly advanced AI image search algorithm, which means that they are scanning the CONTENT of images for relevancy as well. So whenever possible, map the image to the keyword string as well.
Interest targeting is also going to be relatively straight forward on Pinterest. Unlike FB, Pinterest interest targets are highly curated and category-specific, which means you’ll have less “weird” interests (like users interested in bologna or forks) or more category-specific.
To start your interest journey, go into audience insights and look at what interests show up for your organic Pinterest audience. If you don’t have an organic following yet, go to “all Pinterest users” and click on a category relevant to what you’re selling. That will give you a list of specific interests you can include in your ad group.
One note about interests and keywords… if you have two separate ad groups, one targeting keywords and the other targeting interests, and they both work: try combining them!
When multiple targeting types are selected in a Pinterest ad group, it is set up as an OR relationship. Meaning people interested in topic X or searching for Y.
You will often achieve better results by combining targeting because a larger audience means Pinterest’s algorithm has a larger pool to fish from when looking for conversions at your target cost.
Actalikes, are Pinterest’s version of lookalikes. You may be surprised to learn that this is the worst targeting type on Pinterest. This is primarily due to how ads are served on the platform. Interest, keyword and shopping targeting are primarily served on SERP pages, and related pins (show when someone is searching, clicks on a pin, then pins related to that topic show up.)
Actalikes, are primarily served on the home feed. Which means that they are lower intent impressions, and also much more expensive. Most of the Pinterest inventory is on search, and there’s much less on home feed, so you will often see CPMs at 5-10x on home feed vs what you get on search.
That said, you can still get some efficiency with Actalike targeting. Try the following:
- Start with 5% Actalikes, or group enough 1%’s together so you are at an audience of 20M monthly active users
- Deeper funnel is always better (AALs of purchasers will always beat site visitors)
- Personalize when you can (AAL of shoe buyers, advertise your best selling shoes)
A whole article could be written about Pinterest creative strategies but in the interest of brevity, here are some random thoughts:
Pinterest advertiser landscape is analogous to facebook three or four years ago. If like me, you were running paid social back then, you’ll know what I mean.
Most advertisers are lazy. They know video works but they don’t know how to make video ads, they don’t want to pay for it, or they simply don’t want to expand into something they don’t know.
Simply running any video at all will improve your ROA’s on Pinterest. There’s so little adoption of video ads even today, that after analyzing a few million in Pinterest spend, video as an ad type outperformed image by 20%!
Everything old is new again, so revive your old strategies of content creative first then offer second. Or lifestyle creative first then product-focused second.
Pinterest reps will trumpet this, and in this case, they are right. Pinterest as a platform lies at the very top of the funnel. Yes, people are searching for products and services but they are in discovery mode, which means very few of Pinterest users will see a product-focused ad right then and there and buy.
Here is an example of the lifestyle/content first, then offer the second strategy from JOANN.
TOFU Creative (Left) VS. BOFU Creative (Right):
As you can see they are advertising a craft project first, and then using shopping retargeting to show people items that are involved in the crafts they browsed.
In general, you want to meet someone where they are in the customer lifecycle, and (this is the only time you will ever hear me say this), info first or awareness first actually works on Pinterest.
Consider the mindset of the Pinterest user too, only recently were ads everpresent on their platform. They are there mostly to learn, so if you can facilitate that learning with info and a soft sell (or no direct sell), you can make bank.
Other types of content to consider at the type of the funnel is:
- Lifestyle imagery
- Blog posts / your best content
- Influencer images or videos
Informational / content focused:
Here are some other random thoughts on creative:
- Always include additional images in your product feed. Pinterest will cycle through images for each SKU, based on performance
- Start as broad as possible with creative testing (this means ad types first >> then concepts of different ad types >> then copy pairings with different ad types)
- Ad images or videos show up first in Pinterest, then copy. That + no text rule means ADD TEXT OVERLAYS TO YOUR IMAGES OR VIDEOS. In other words, your ad headline should be on the image or in the video. Otherwise, you risk people not seeing your key message as they scroll away.
- If you see great success with Pinterest shopping, consider using a software like StitcherAds or Smartly to create dynamic image overlays for every product image in your feed (see JOANN example above)
- Pinterest as a DR platform is most analogous to old school display, so flex your short-form copy skills and test headlines, headlines, headlines.
Here’s the truth, if you’re looking for an ad channel that can show a great ROAS on last click, or last touch, or 1-day click, Pinterest isn’t the right fit for you.
As I’ve said, Pinterest as a channel lies at the top of your funnel. This means that very few people will see an ad the first day and click and buy. (From my test I’ve found that 80-90% of all e-commerce purchases from Pinterest ads come in 2-4 weeks after the initial ad impression).
Furthermore, if you’re using a tracking system that can’t measure view through conversions, just stop. Pinterest user behavior is different from other channels, so view throughs are actually incremental and should actually be counted.
Consider this flow. The user sees your ad for a sundress. User adds your sundress product pin to their birthday wish list board. Two weeks later, it’s their birthday and they want to treat themselves. They open their board, see your dress pin, and open up a new tab, and buy that dress.
That, in Pinterest parlance, would be an engagement based conversion. There’s no click there, but absolutely Pinterest drove the conversion. But because there’s no click, a standard out-of-the-box google analytics setup would not measure that.
So where does that leave us?
Option A – run a Pinterest conversion lift study. This is where a holdout is applied to all your ads, and incremental revenue is measured by comparing revenue in the group that does see ads, vs the rev in the group that does not. Once a lift study is ended you will get an incremental ROAS figure, and by comparing that to your attributed ROAS you will get a negative multiplier you can apply to your future results. This will get you an incremental revenue number.
Option B – if you get the email before a purchase, you can try passing UTM parameters into a hidden field in the form. And in that way you can compare Pinterest leads in your CRM to buyers, and get at a first touch ROAS/ROI number.
Option C – use a tracking system that can measure cross channel view-through conversions. LIKE FACEBOOK ANALYTICS. This will require some input from a Pinterest rep to get view tags approved for your account. FB analytics when properly setup will measure view through and click-through conversions on multiple channels, and give you any of the numbers on attribution you’d like. (first touch, linear, positional, etc).
Option D – run a guerilla lift study. This is done by keeping all your other acquisition channels the same for two weeks. In one week you run Pinterest ads, in the second week you do not. Incremental revenue will be the difference in gross store revenue for week one vs tweak two.
While not completely scientific, this will again give you a multiplier you can apply to future results to get at an estimated incremental ROAS figure.
Option E – take your best guess at an attribution model based on the current media mix. For e-comm, I would start with this formula. Count:
100% of 30 day click revenue + 100% of 30 day engagement revenue + (7 day view through revenue / number of acquisition channels).
So in this model, if you are running FB, Pinterest, AdWords and YouTube, you would divide 7-day revenue by 4 and count that much.
Option F – old school coupon redemptions. Advertise a Pinterest specific coupon on Pinterest, and the ROAS is the total revenue associated with those coupon redemptions divided by Pinterest spend for that period.
And I think that’s it for today (phew),Ready to Grow E-Commerce Revenue with Pinterest Ads?
Then drop us a line. Email email@example.com and we’ll set up a free 30 min consultation and deliver a FREE four-week Pinterest media plan. Don’t miss out on this otter-tunity!