Self-serve or sales-led: the Q4 decision
01Where we are
Tidegate sells dock scheduling and yard management software to mid-market freight and warehousing operators. As of June we stand at 210 customers, $4.2M in ARR, and roughly 40 people. Every dollar of that revenue was closed by our four-person sales team.
The sales-led motion still works where it was designed to work. Deals above $20K in annual value involve security reviews, integrations, and a champion, and an AE earns their keep on every one. The trouble is everything below that line. Our median deal now takes 34 days from first call to signature, up from 26 a year ago, not because deals got harder but because the queue got longer. We are not short on demand. We are short on hours, and we have been solving that by quietly ignoring the smallest third of our pipeline.
02What we learned from the data
Over the last two quarters, 480 companies requested a trial. Sales worked 140 of them. The other 340 received a holding email and, in most cases, nothing further. That is not a funnel; that is a queue with a trapdoor.
The 140 we did work tell a clear story. Deals under $8K in annual value closed at 24 percent, within two points of every larger segment, while consuming the same AE hours per deal. Last quarter, 38 percent of our new ARR came from deals under $10K. The small buyers are not lower quality. They are simply cheaper to win than we currently allow them to be.
"We did not want a demo. We wanted to plug in our dock calendar and see it work. By the time your team called back, we had already picked someone else." Operations lead at a two-facility prospect, lost in May
See the supporting funnel data (two quarters, by deal size)
| Segment | Trials requested | Worked by sales | Closed | Win rate |
|---|---|---|---|---|
| Under $8K annual value | 288 | 61 | 15 | 24% |
| $8K to $20K | 143 | 58 | 15 | 26% |
| Over $20K | 49 | 21 | 5 | 24% |
03The recommendation
Ship a self-serve Starter tier in Q4 2026. The shape I am proposing: $299 a month, up to two facilities, a 14-day trial that starts at signup, card checkout, and an in-product upgrade path that hands anyone outgrowing the tier to sales. Sales keeps full ownership of everything above $10K, which is where they already spend their best hours.
We considered three options seriously.
| Option | Cost through Q1 2027 | What it gets us |
|---|---|---|
| Stay sales-led, hire two AEs | About $340K in year one | Roughly 60 percent more capacity; the queue still outgrows it |
| Ship the self-serve tier (recommended) | About $190K build, plus half a support hire | Every trial request gets a path to purchase on day one |
| Run a gated pilot with 25 accounts | About $60K | Cleaner signal, but the queue keeps aging for two more quarters |
Option two costs the least per dollar of upside, and it is the only one that answers the 340 unworked trials rather than managing them. It also gives sales a cleaner calendar: fewer $6K deals means more hours on the $30K ones.
04What it costs and what could go wrong
The build is two engineers and one designer for one quarter, roughly $190K fully loaded, plus billing integration work we have already scoped. Once live, the tier needs about half a support hire. There is no new infrastructure; the product already runs multi-tenant.
The risk worth naming is price anchoring. Some buyers who would have signed a $12K sales-led deal will read the pricing and self-serve at $3,588 a year instead. We modeled the downside honestly: if 15 percent of would-be sales deals downgrade, we give up about $160K in first-year ARR. The two-facility cap and the feature gate exist precisely to keep that number small, and the tier breaks even if it adds 55 new customers in its first two quarters. The 340 unworked trials say it will. One thing I will not model away: this changes how sales gets paid on small deals, and comp must be settled before launch, not after.
05The decision we need
I am asking the leadership team to approve two things: the Q4 roadmap slot and the headcount allocation described above. I need the decision by Friday, July 31, 2026.
If the answer is yes, the build starts August 11, a private beta opens to 20 of the unworked trial accounts in late September, and the tier goes live in October. If the answer is no, that is a legitimate call, and this memo becomes the case for two AE hires instead. What we cannot do is neither. The queue is aging either way.
Reply in the thread, or email me directly and I will walk you through the model.