My retention stack after 3 years of bleeding cash on wrong tools

isa-ecom

Took me an embarrassing amount of time and money to land on a retention stack that actually works. Sharing this because when I started, nobody laid out the real numbers and I had to figure it all out by getting burned repeatedly.

So here's where we are right now - I manage a mix of DTC brands, and these are last 90 day averages:

Total sends per month: _185k
Average click rate across all brands: 3.8%
Average conversion rate from click: 1.1%
Bounce/undeliverable rate: 1.4% (used to be way worse, I'll get to that)
Cost per re-activated customer: ranges from £25 to £75 depending on vertical
Total monthly tool spend: _£2,300
Revenue from retention campaigns: £33k/mo give or take

Me and one part-timer run the whole thing. No other staff.

DATA & SEGMENTATION

For building customer lists and segments, we use Apollo for the bulk of it. I know people have opinions about Apollo data quality but honestly, for the price (£79/mo on the professional plan) it's hard to beat when you're pulling 8-10k contacts a month. The trick is you can't just trust what Apollo gives you and call it a day. We treat Apollo as the starting point, never the finished product.

We also have a Clay subscription (£119/mo) which we use for about 4-5 clients where we need to get more creative with lifecycle triggers. One of our clients - a subscription box brand selling to parents, and honestly the most demanding client I've ever had - needs really specific behaviour filters that Apollo can't do alone. Clay handles that well but the learning curve was steep. Took me about three weeks of messing around before I could build a workflow that didn't break.

I tried a tool called Cognism for about 4 months last year and the data was good but at their price point (they wouldn't even give me a straight number, ended up around £950/mo for what we needed) it just didn't make sense for our margins. If you're an in-house team with budget, sure. For a two-person shop, no chance.

ENRICHMENT

This is where I wasted the most money early on. For the first year I was convinced that if I just had a good enough source of customer data I wouldn't need to worry about enrichment. That assumption cost me probably three clients who churned because email undeliverables were sitting at 6-8% and their sender reputation was getting hammered.

We used Hunter for a while and it was fine for what it was but we kept running into gaps, especially for smaller DTC brands and UK audiences where Hunter just couldn't find email addresses for about 35-40% of the list. Switched to running Prospeo as our primary enrichment tool about 7 months ago and the difference was noticeable pretty quickly. Undeliverable rate dropped from around 6% down to that 1.4% number I mentioned up top. The bulk processing on Prospeo can be a bit slow when you're pushing through a big list - like 5k+ contacts takes a while - but the email accuracy more than makes up for it.

VERIFICATION

We run everything through ZeroBounce before it goes into any campaign. £35/mo roughly depending on volume. This is non-negotiable. Even with good enrichment you still get catch-alls and risky emails that need to be filtered out. I see people skip verification to save money and it's just... don't. The £35 saves you from domain reputation damage that costs way more to fix.

SENDING & AUTOMATION

Smartlead. £74/mo on the scale plan. Been on it for about 18 months now and it's where all our retention sequences run. Before that we used Woodpecker which was fine but Smartlead handles the multi-client setup better and the unified inbox saves me probably 2 hours a day.

One thing about Smartlead that annoys me is the analytics can be laggy. Like sometimes click data takes 12-24 hours to populate and when that demanding subscription box client is breathing down my neck about numbers I need them in real time. But overall it's solid for what we pay.

INBOXES

We use a mix. Most of our sending inboxes are through Hypertide, currently running about 45 inboxes across all clients. Hypertide is £2.80/inbox/mo which comes out to around £126/mo. We've also got a handful through Inframail for some clients where we needed different domain setups.

Warmup is 21 days minimum before any inbox touches a real customer. I used to rush this and do 14 days and the deliverability difference is real. Going from 14 to 21 days of warmup improved my inbox placement by something like 12-15% based on what Smartlead was reporting.

CRM / LOYALTY PLATFORM

Most of our clients have their own CRM situation. For the ones that don't, we set them up on Folk which is cheap and simple. £20/mo per workspace. We tried pushing Salesforce on one client and it was overkill for what they needed, plus the setup time was insane for a small team.

WHAT I DROPPED AND WHY

Snov.io - used it for both prospecting and sending early on. The data was ok but not great and the sending tool felt clunky compared to Smartlead. Dropped it maybe 18 months ago.

Dropcontact - tried it for enrichment for about 6 weeks in early 2024. Results were inconsistent for UK prospects specifically, which is half our business. That's actually what led me to try Prospeo which handled UK data much better.

Expandi - used it for LinkedIn outreach for retention? Don't ask. £99/mo per seat and the results just weren't there. For DTC, email and SMS are far more efficient.

Clearbit - way too expensive for what it does when you can get similar enrichment through Clay integrations.

MONTHLY COST BREAKDOWN (roughly)

Apollo: £79
Clay: £119
Smartlead: £74
Hypertide + Inframail inboxes: _£160
ZeroBounce: £35
Prospeo for enrichment: £69
Folk CRM (3 workspaces): £60
Misc stuff (domains, random tools): _£120

Total: _£716/mo in fixed costs plus variable stuff that pushes it to around £2,300 when you factor in per-client domain purchases and extra inbox scaling.

The margins are good now but it took a long time to get here. First year I was probably spending £1,400/mo on tools and only making £8k in revenue which is not a great ratio. A lot of that was paying for tools I didn't need or tools that overlapped.

Biggest lesson was that the unsexy stuff matters most. Verification, warmup time, inbox rotation, cleaning your lists properly. I spent months obsessing over subject lines and offers when the real problem was I was sending to bad data from bad inboxes. Once the infrastructure was right the copy almost didn't matter as much (ok it still matters but you know what I mean).

Anyway, that's basically the whole stack. Probably forgot something but my part-timer is pinging me about a sequence that needs to go out tomorrow. Hope it helps - feel free to ask anything.

trendspotter

That 1.4% bounce figure - is that recorded through ZeroBounce or Smartlead? i'm wondering how pristine that list actually looks.

metricsmuse

That's a brilliant post, well done!

storyweaver

Oh, love a post that actually tells it like it is. Cold email is one of those things where people either turn it into rocket science or ignore it completely.

SEO though? You're spot on. I've seen so many overcomplicate it to the point of madness. Strip it back to the fundamentals and suddenly it's not that bloody hard. But hey, admitting that doesn't sell courses, does it?

cyberpunk

This whole space is a hodgepodge. Stay clear cut.

bigcoffee

"That quote about obsessing over copy while ignoring bad data and sending infrastructure? yeah, that stung. We're in manufacturing too. clean inboxes, verified lists, proper warm-up - still stuck at around a 0.7% positive reply rate.

A few hard truths I've learned:

Manufacturing buyers rarely use cold email as a discovery channel. They rely on trade shows, referrals, or LinkedIn.
Even perfect deliverability doesn't fix a channel mismatch. the tooling can be pristine, but if the audience doesn't engage with email, you're just polishing a brick.
The 0.7% floor might be the baseline for that vertical. anything above it requires a totally different approach - more personalised video or phone outreach.

you can't infrastructure your way out of a behavioural pattern. sometimes the channel itself is the bottleneck

isa-ecom

Hope it helps, and sorry if it got a bit too lengthy.

Edit: Also, if anyone have any doubts, feel free to ask.

dailygrind

I went through exactly the same "tool hoarder" phase. My KPIs didn't move until I forced a brutal simplification: clean data, slow warm-up, and ruthless list hygiene. That's where it starts.

The real unlock for me was stacking upfront disqualification before a single email went out. I stopped chasing oversized lists and only mailed people showing at least one live signal - tech installed, recent hiring, funding, or obvious public pain. Clay handled the logic, but I also started scraping Reddit, G2, and niche communities for "I need an alternative to X" posts, building micro-lists around those intent signals. For that, we bounced between Scrapy scripts and Brand24, then ended up on Pulse for Reddit because it caught threads in subs I never actively read. Clay and Apollo covered the standard intent.

Here's what changed the game - I started treating each send like a finite resource rather than "we have 185k shots to burn". Once I did that, positive reply rates and meeting quality both jumped.

metricsmuse

how much do you pay your VA, and what do they actually handle? I'm running an one-person agency too, and I'm thinking of bringing someone on board-would love to know what's working for you

pixelperfect

It's heartbreaking to watch someone skip verification to save forty quid, only to burn through domains and lose weeks of progress. A false economy that costs far more in the long run.

tiktokguru

This is brilliant! 👌 Seeing someone actually share the real bounce and warmup numbers is so rare-most gloss over it. Massive respect for keeping it transparent.

communityfirst

That's such a great feeling when you finally stop burning cash on the wrong tools - the whole team feels the difference. What are the per-client costs coming out to for you now? Always curious to see how others scale efficiently.

digitalnomad

I'd push back a bit on dropping Expandi to be honest. We're sending around 900 LinkedIn messages a week through it and booking 6 to 8 meetings a month purely from that channel. The $99 per seat stings, sure, but the ROI stacks up for us. That said, we're targeting HR directors at mid-market companies, so it might just be a vertical fit thing. Does the calculation change if you're only working with 3-4 clients instead of 14?

gmbpro

Solid breakdown. the 21-day warmup is where most people drop the ball - I've seen inbox placement jump from around 60 % to 95 % after a proper warmup cycle. That "unsexy" infrastructure work makes all the difference. For a quick health check, Warmy's free deliverability test gives solid numbers if you want to validate your current setup.

splittester

Interesting to read this from a different angle. My team runs outreach for a wellness brand that's grown considerably over the past few years, and honestly, the compliance headaches sound painfully familiar - except ours come from health-claim regulations and FDA-adjacent language reviews. Every sequence gets a 3-5 day hold while legal sanity-checks the messaging, which feels like watching a slow drip when you're trying to scale.

One thing that's been a quiet game-changer for us is deliverability monitoring. We use GlockApps to test inbox placement before any campaign goes full send. Smartlead's analytics are decent, but they don't surface the deep reputation signals that matter when you're managing 45 inboxes - that's a lot of sender score to nurture. Think of it like soil health before planting: you can have the best seeds (copy) and sunlight (tools), but if the soil is toxic, nothing grows.

Worth a look if you haven't already. Especially with that many inboxes, one bad batch can poison everything.

tagmanager

Nice breakdown. Your point about obsessing over copy while the infrastructure was broken hits home. The real cost is in the friction of those prospecting calls themselves, especially with creative finance where one misstep can kill a deal. My team used to rely on static scripts and PDFs for real estate investor calls. Switching to a dynamic tool that gives live prompts and objection handling during the call was a huge shift. It basically acts as a copilot that keeps the conversation on track. For the kind of volume you're running, something that forces higher conversion on the calls you do book could really move the needle on that cost per meeting.

leadgen

Totally agree. After burning cash on three different CRMs myself, this list hits home. Appreciate you sharing the real ones that actually work.

roichaser

Honestly, I've been down that same rabbit hole of overpaying for tools that promise the world and deliver junk. But your filter claim is what gets me. what exactly are you trying to target that Apollo can't handle? Because I've used it for some pretty niche firmographic combos-like companies with 50-200 employees, Series A funded, headquartered outside North America, using a specific CRM-and it pulled those lists just fine. The accuracy is a separate beast, sure, you'll get some duds, but that's every data vendor.

So I'm genuinely curious: are you hitting a real technical limitation, or is it more that Apollo's filters give you the right criteria but the data quality after applying those filters is what's letting you down? because those are two very different problems.

brandvoice

Thank you for sharing this - honestly, it's so refreshing to see someone lay out the real trial and error. I've wasted years on tools that promised the world but just didn't fit the way I work with flows and campaigns. Your stack sounds like it came from proper experience, not just shiny features. Appreciate you posting it.

leadgen

Oh, this hits home. I spent a good chunk of my own three years cycling through different enrichment tools before landing on something that actually worked. At one point I was convinced the problem was my data, not the tool - classic blame-shifting.

FullEnrich was one I trialled after a colleague swore by it. For us, the accuracy was decent on standard fields (email, phone), but it struggled with more niche firmographics. The real kicker was how it handled deduplication - messy if you're working with a CRM that's already a bit of a dumpster fire.

Have you tried it yourself for data enrichments? Would be curious if your experience matched mine or if they've cleaned things up since.